key: cord-0889640-0hmnhilt authors: Miranda, M.T.; Romero, P.; Valero-Amaro, V.; Arranz, J.I.; Montero, I. title: Ventilation conditions and their influence on thermal comfort in examination classrooms in times of COVID-19. A case study in a Spanish area with Mediterranean climate date: 2021-12-23 journal: Int J Hyg Environ Health DOI: 10.1016/j.ijheh.2021.113910 sha: 53daed7f0a05554e07418b6530c5815de92e2fcd doc_id: 889640 cord_uid: 0hmnhilt Current evidence and recent publications have led to the recognition that aerosol-borne transmission of COVID-19 is possible in indoor areas such as educational centers. A crucial measure to reduce the risk of infection in high occupancy indoors is ventilation. In this global pandemic context of SARS-CoV-2 virus infection, a study has been carried out with the main objective of analyzing the effects of natural ventilation conditions through windows on indoor air quality and thermal comfort during on-site examinations in higher education centers during the winter season, as this implies situations of unusual occupation and the impossibility in many cases of taking breaks or leaving classrooms, as well as the existence of unfavorable outdoor weather conditions in terms of low temperatures. For this purpose, in situ measurements of the environmental variables were taken during different evaluation tests. As the main results of the study, ventilation conditions were generally adequate in all the tests carried out, regardless of the ventilation strategy used, with average CO(2) concentration levels of between 450 and 670 ppm. The maximum CO(2) concentration value recorded in one of the tests was 808 ppm. On this basis, the limit for category IDA 2 buildings, corresponding to educational establishments, was not exceeded in any case. However, these measures affected the thermal comfort of the occupants, especially when the outside temperature was below 6 °C, with a dissatisfaction rate of between 25 and 72%. Examinations carried out with outside temperatures above 12 °C were conducted in acceptable comfort conditions regardless of outside air supply and classroom occupancy. In these cases, the dissatisfaction rate was less than 10%. The results obtained have made it possible to establish strategies for ventilation in the implementation of future exams, depending on the climatic conditions outside. or they can be inhaled directly by others if they are close enough (Faridi et al., 2020) . However, as mentioned above, there is evidence of transmission by aerosols of 100 μm or less in size, which can infect people at a distance of more than two meters. These transmissions usually occur in enclosed spaces with inadequate ventilation, where people stay for a long period of time (CDC, 2020; WHO, 2020a) . A crucial preventive measure to reduce the risk of contagion in indoor spaces, in addition to the use of masks, interpersonal distance and hygiene measures, is ventilation, defined as the renewal of indoor air with outdoor air, either by natural or mechanical means or a combination of the two systems (Ministerio de Sanidad and Gobierno de España, 2020). On the one hand, natural ventilation is achieved by non-mechanical means, normally by opening doors and windows, taking advantage of the pressure differences generated by a temperature gradient or by the action of the wind. The highest efficiency is achieved with natural cross ventilation, in other words, by opening two doors or windows in opposite walls of the room to promote air circulation and ensure an efficient sweep throughout the space. In situations of high COVID-19 transmission, the prioritisation of natural ventilation should be assessed against the thermo-hygrometric conditions necessary for thermal comfort or energy efficiency requirements (Atkinson and WHO, 2009 ). Mechanical ventilation controls the air inlets and outlets, so it is not influenced by outdoor weather conditions and allows control of the flow rate introduced. It is recommended that, even if there is mechanical ventilation, natural ventilation is regularly carried out by opening doors and windows in order to achieve good combined ventilation (INSHT, 2006a) . Air ventilation rate per hour is used to check the air renewal in a given place (ACH) (Ministerio de Sanidad and Gobierno de España, 2020). On the other hand, to evaluate indoor air quality (IAQ) as well as ventilation conditions, the level of CO2 concentration is used, as this is a good indicator of human bioeffluent emissions (INSST et al., 2021) . Educational establishments are environments vulnerable to the transmission of the SARS-CoV-2 virus regardless their level, as they are enclosed spaces where the activities carried out involve a large number of contacts in which a safety distance cannot be kept, so preventive measures such as the use of masks and ventilation are essential to curb possible infection. There are studies that associate poor quality classroom environments with an increased risk of respiratory and allergic diseases, as well as compromised academic performance of students and working conditions of teachers (Baloch et al., 2020; Grineski et al., 2016) . Minguillón et al. have developed guidelines for ventilation in nursery and primary school classrooms, in which they set out different measures to reduce the risk of infection. This document set out recommendations for effective ventilation J o u r n a l P r e -p r o o f and air purification based on room volume, number and age of occupants and activity. In addition, it provided tools to determine whether the ventilation conditions achieved are adequate (Minguillón et al., 2020) . In the same line, Allen et al. carried out a manual to measure the air renewal rate in classrooms (Allen et al., 2020) . Villanueva et al. studied ventilation conditions and particulate matter in 19 childhood education, primary and secondary school classrooms located in the metropolitan area of Ciudad Real (Spain) during the reopening of schools after confinement. This study showed that preschool classrooms were the educational environments with the lowest average CO2 levels (553 ppm), while secondary school classrooms had the highest average carbon dioxide concentration, with values close to 700 ppm (Villanueva et al., 2021) . Similar work has been done by Zemitis et al. in secondary school in south-eartern Latvia. In this study, CO2 concentration levels significantly exceeding 1,000 ppm were obtained in all the classrooms studied. The average concentration was approximately 2,380 ppm and even reached absolute maximum levels of 4,424 ppm (Zemitis et al., 2021) . Vassella et al. developed an intervention study to verify whether the recommended indoor air quality objectives can be achieved by following reasonable ventilation regimes that are also suitable for countries with cold winters. To this end, they first measured CO2 levels in classrooms without any ventilation intervention and compared the effectiveness of natural ventilation during breaks only in 100 primary and secondary classrooms. It was found that the average CO2 levels were reduced from 1,600 ppm to 1,097 ppm, demonstrating the effectiveness of ventilation in trying to control possible COVID-19 infection (Vassella et al., 2021) . Finally, Asif and Zeeshan monitored and assessed indoor CO2 levels in naturally ventilated classrooms in Pakistan, among other parameters. It was found that carbon dioxide concentrations exceeded those recommended by ASHRAE, even reaching values above 4,000 ppm when the classrooms were occupied (Asif and Zeeshan, 2020) . Requirements for increased natural ventilation mean that thermal comfort conditions inside classrooms can be altered, especially when there are low temperatures outside, as is the case in the winter season. Environmental conditions assessment in schools has been addressed by different authors. A study by Alonso et al. analysed the effects of the COVID-19 pandemic on thermal comfort and IAQ in winter. For this purpose, in situ measurements of environmental variables were carried out before and during the pandemic in two classrooms of a primary school located in southern Spain. The results showed a reduction of 400 ppm when the schools were naturally ventilated during all teaching hours. However, the analysis of standards shows that over 60% of hours are thermal discomfort conditions (Alonso et al., 2021) . Heracleous and Michael evaluated the impact of natural ventilation on the indoor thermal environment through an extensive study conducted in both winter and summer in schools located in Mediterranean climate zones. For the winter period, students felt neutral or slightly cool, with mean wind chill values of -0.07 and a percentage of dissatisfied students J o u r n a l P r e -p r o o f of approximately 30% (Heracleous and Michael, 2020) . In a study carried out in primary schools in the northern part of Sweden during the heating period, parameters related to thermal comfort were measured and the Predicted Mean Vote (PMV) was calculated (Yang et al., 2018) . Finally, Jiang et al. developed a thermal comfort assessment model emulating different ventilation conditions in primary and secondary schools in rural China during winter. The result subsequently indicated that the comfortable temperature ranges for 90% of the pupils was 13-18˚C (Jiang et al., 2020) . In terms of studies analysing thermal comfort in settings other than educational establishments. For example, Pourshaghaghy and Omidvari evaluated the level of thermal comfort in several areas of a hospital in Iran. PPD values were higher than 10% in all areas of the building, with the worst thermal conditions in the surgery section (Pourshaghaghy and Omidvari, 2012) . On the other hand, in hot climate zones, the use of techniques to cool the air is key to providing an acceptable level of comfort. Therefore, Yüksel et al. investigated how these measures affect a mosque in Turkey. The use of air conditioning improved the overall thermal sensation from a PPD of 40% to 13% (Yüksel et al., 2020) . Most of the studies carried out so far in this field focus exclusively on the assessment of ventilation conditions when teaching in pre-school, primary and secondary schools. However, these studies do not take into account thermal comfort conditions for teachers and students, which could be significantly affected by an increase in the intake of colder air during ventilation in winter. Nor do they take into account special conditions such as exams, where, as a general rule, they take longer than theoretical and practical classes and, breaks and departure are not possible when necessary. On the other hand, university classrooms where teaching takes place are considerably more occupied and there are laboratory practices where measures to reduce the risk of infection may be more complicated to carry out. Also, depending on the age of the university students, specific CO2 generation rates per person are required that have not been considered in other studies. Based on the above, a study has been carried out to assess the indoor air quality and environmental conditions existing during on-site examinations in a higher education center located in the southwest of Spain during the January 2021 exams. For this purpose, ventilation conditions have been studied, establishing different strategies by measuring in situ the existing CO2 concentration levels depending on the same and the number of occupants present in each classroom. Similarly, the ventilation rate and external air flow rate have been determined. It has also been examined how the measures adopted to try to avoid COVID-19 infections interfere with the thermal comfort of the occupants depending on their clothing and the type of activity carried out in the classrooms. The results obtained have allowed us to discern the most effective ventilation strategies in order to maximize thermal comfort inside the classrooms. Official closing date for university centres in Spain by COVID-19 took place on the 16th of March 2020. For the return to on-site university education scheduled for September of the same year, a contingency plan was established for educational places that included measures in strict compliance with existing WHO recommendations for infection control, mainly promoting frequent hand hygiene, ensuring regular cleaning of surfaces, maximising physical distance to maintain at least 1.5 meters of interpersonal distance and increasing ventilation of indoor spaces. In addition, the use of face masks by students and teachers was obligatory throughout the entire period in all areas of educational centres (WHO, 2020b) . It was also agreed that presential exams would be held at the universities during the 2020-2021 academic year. For the assessment of IAQ and thermal comfort, a university centre located in Extremadura, a region in the southwest of Spain, was selected (38˚53'2.5''N, 7˚00'11''), located in an area with a Mediterranean climate. Fig. 1 shows the location of the higher education institution under study, while Fig. 2 shows the location of each of the classrooms studied. Additional elements such as photographs of some of the classrooms and the north facade of the building are also included in this figure. The January exam period, which ran from 11th to 29th January 2021, was chosen for the study because of the lowest outdoor temperatures during the winter season. Specifically, during the first half of the selected period, temperatures were notably colder due to the presence of the Filomena squall over the Iberian Peninsula . During this period, moreover, the 14-day cumulative incidence of COVID-19 in the area where the study was conducted was around 1,500 cases, while in Spain it was approximately 890 cases per 100,000 inhabitants. This meant that all tests had to be carried out with the maximum possible ventilation as the situation was considered to be extremely risky (CCAES et al., 2021) . A total of 88 exams, some of them with several groups, were held in a total of 13 classrooms. All classrooms in the selected sample had exclusively natural ventilation. Windows and doors remained open at all times in most examinations, except for those where a different ventilation protocol was established. In one case, windows were opened according to the CO2 measurements that were recorded, and in other cases, windows were partially opened, specifically windows in front of the door. Based on the above, the following criteria were taken into account for the selection of classrooms and examinations sampled. As a main criterion, classrooms where ventilation strategies other than the total opening of doors and windows were carried out were considered. Classrooms were also chosen where a higher number of exams were held and which were located on different floors of the building. On the other hand, examinations were considered for all the degrees taught at the university centre under study and for all academic years, as these have a variable number of students. Thus, a total of 7 classrooms (53.8%) out of the 13 that were available for the exams held in January were analyzed (unselected classrooms have identical characteristics in terms of size and number of windows). With regard to the number J o u r n a l P r e -p r o o f of examinations, measurements were carried out in 18 of the 88 examinations planned (20.5%), of which 10 were in the morning and 8 in the afternoon. Different lengths of exams were also considered, some of them including a break. It should be noted that, in morning exams, windows were opened 30 minutes before the exams started and remained open until the end of the afternoon exams. To evaluate the ventilation conditions in different classrooms during the tests, as well as to monitor the IAQ, the level of CO2 concentration (in parts per million, ppm) was used as an indicator. In this study, a portable instrument was used to measure and record the existing CO2 concentration (model PCE-AQD 20, PCE Instruments, PCE Deutschland GmbH, Meschede, Germany). CO2 measurement range was from 0 to 10,000 ppm, with a resolution of 1 ppm. Equipment accuracy is as follows: <1,000 ppm: ±40 ppm, <3,000 ppm: ± (50ppm + 3% of the value), >3,000 ppm: ± (50ppm + 5% of the value). CO2 concentration levels were measured and recorded at an interval of 2 seconds. At the same time, a thermal environment meter (model HD32.1, DeltaOHM, GHM Group, Remscheid, Germany) was used to assess the microclimate in the classrooms during the exams. In this case, a combined probe for measuring air temperature and relative humidity, with a measuring range of -40˚C to +100˚C; a probe for measurement of balloon temperature, with a measuring range of -10˚C to +100˚C; a sonde for radiant temperature recording, with a measuring range of 0˚C to +60˚C; and an omnidirectional hot wire probe for air velocity recording, with a measuring range of 0˚C to +80˚C. The atmospheric pressure measurement accuracy is ±0.5 Pa, with a response time of 1 Hz. Temperature measurement range of the instrument is -200˚C to +650˚C, with a resolution of 0.01˚C in the range ±199.99˚C and 0.1˚C and an accuracy of ±0.01˚C in the range ±199.99˚C and ±0.1˚C in the remaining field. Finally, the relative humidity measuring range (capacitive sensor) is 0 to 100% RH, with a resolution of 0.1% RH and an accuracy of ±0.1% RH. DeltaLog 10 software associated with the device was used for data reading and processing. Continuous monitoring of environmental parameters and CO2 concentration of the classrooms was carried out throughout the duration of the selected exams. CO2 measuring equipment was placed at the end of the classrooms, in order to avoid direct exhalation by the occupants (Heracleous and Michael, 2020) , at a height that coincided with breathing zones of the occupants, keeping a distance of at least 1.5 meters from walls and at least one meter from people (Griffiths and Eftekhari, 2008; WHO, 2020a) . Thermal environment measuring equipment was placed in the central part of the rooms in order to representatively assess the microclimate present during tests (AENOR, 2002; Yang et al., 2018) . Both equipments were set up and activated 15 minutes before the students entered the classroom for exams to ensure the steady state of the measurements. Air temperature, relative humidity and outdoor CO2 concentration were measured J o u r n a l P r e -p r o o f before and after each test. Data on outdoor air speed were obtained from the Spanish State Meteorological Agency (Agencia Estatal de Meteorología, AEMET) (AEMET, 2021). The university center under study has a built surface area of 13,055 m 2 and a usable surface area of 11,418 m 2 , distributed over 4 floors. All classrooms in this building only had natural ventilation. Interior partitions between classrooms are made of concrete blocks with a direct plasterboard backing. Building has a hot water radiator heating system, which was kept on during the sampling. Fig. 3 shows the geometry of each of the classrooms studied, as well as the location of the measuring equipment used. All the windows are shown in blue. All classrooms are rectangular shape and have openings for cross ventilation, described as the best approach to natural ventilation. Slight differences in the dimensions, shape and configuration of openings may influence the ventilation efficiency (Chenvidyakarn and Woods, 2005; Lee and Choi, 2002) , but since all classrooms are cross-ventilated and similar in shape, this aspect was not considered in the work. Windows face north, opening to the outside where there is a one-storey building in the immediate vicinity. They have two leaves of 1. Classroom occupancy ranged from 5 to 40 people (teachers + students). Therefore, the occupancy/volume ratio was between 0.01 and 0.10 persons/m 3 . In other studies, such as the one carried out in Ciudad Real (Spain), average occupancy was 24 people in classrooms of approximately 60 m 2 (Villanueva et al., 2021) , values significantly higher than in this study (with minimum ratios of 3.4 m 2 /person and maximum of 22.5 m 2 /person). Classrooms studied in Switzerland had a volume of 200 m 3 , also smaller than almost all the classrooms analyzed, and the window area was 5 m 2 , a higher value than the one available in this study (Vassella et al., 2021) . In the work by Asif and Zeeshan, classrooms' areas studied were significantly smaller J o u r n a l P r e -p r o o f (from 20.3 to 33.9 m 2 ) than those analyzed in this paper, and the open area of windows was also smaller. However, the average occupancy was around 20 people, so the stocking density was 1 m 2 /person, slightly more than three times lower than the minimum in this work (Asif and Zeeshan, 2020) . J o u r n a l P r e -p r o o f In Spain, the UNE-EN 13779:2008 standard defined the requirements for ventilation and air-conditioning systems in non-residential buildings. This standard was the document on which the Regulation on Thermal Installations in Buildings (RITE) was based for this type of installation (AENOR, 2008) . RITE classifies indoor air quality (IAQ) in four categories (IDA, Indoor Air), depending on the use of buildings, proposing in each case an outdoor air flow rate per person. Classrooms in educational establishments belong to IDA category 2, therefore an outdoor airflow of 12.5 l/s per person is recommended (European Parliament, 2002; RITE, 2007) . This value reflects good air quality. However, to reduce the risk of COVID-19 infection, an outdoor air flow of 14 l/s per person per second is recommended (Ministerio de Sanidad and Gobierno de España, 2020). Both ASHRAE in ASI/ASHRAE 62.1-2019 and the Harvard Guide recommend a rate of 5-6 ACH for classrooms between 80 and 100 m 2 and 25 students (Allen et al., 2020; ASHRAE, 2019). As mentioned above, one measure used as an indicator of IAQ is to use the CO2 concentration level. For each indoor air quality category, the RITE sets permissible limit values for carbon dioxide concentration (in ppm) above the CO2 concentration in outdoor air. For category IDA 2 it is recommended not to exceed a carbon dioxide concentration of 500 ppm above the CO2 levels in the outdoor air (AENOR, 2008) . A carbon dioxide concentration corresponding to an IDA 1 category establishes an optimal indoor air quality that is obtained with CO2 concentration values below 350 ppm (measured over the outdoor CO2 concentration). On the other hand, CO2 levels above 800-1000 ppm (measured over the outdoor CO2 concentration) could be an indicator of poor indoor ventilation. However, this CO2 concentration is far from being harmful to human health and should only be interpreted as an indicator of the need for ventilation (Ministerio de Sanidad and Gobierno de España, 2020). Table 2 shows recommended values for the two measures used in the ventilation analysis in terms of indoor air classification (IDA), that is, outdoor air flow per person, and CO2 concentration. , 2020, 2006) . Percentage values of dissatisfied people up to 10% reflect a satisfactory situation for the majority of people (90% satisfied), whereas higher values indicate a situation of thermal discomfort. This PPD value corresponds to the range between -0.5 and 0.5 indicated for the PMV (Fanger, 1970) . Table 3 shows the comfort limits for each of the categories. Ventilation rate was estimated by the tracer gas method using CO2 generated by the occupants themselves as tracer gas (Batterman, 2017; Remion et al., 2019; Schibuola et al., 2016) . This method is based on the mass balance equation according to which the change in the amount of tracer gas present is obtained by the difference between that generated plus that introduced and that eliminated, as shown in the following differential equation (Almeida and De Freitas, 2014; Griffiths and Eftekhari, 2008; Krawczyk et al., 2016) . J o u r n a l P r e -p r o o f Where V is classroom volume (m 3 ), C(t) is the concentration of the tracer gas at time t (ppm), t is time (s), G is the tracer gas generation rate (l/s), Q is the exchange rate between inside and outside, that is, the actual renewals occurring (m 3 /s), and Cex is the outside concentration of the tracer gas (ppm). Carbon dioxide density to air is 1.53 and the reference value is 5,000 ppm (AENOR, 2008). CO2 generation rate per person (in l/s) was calculated by considering the volume of carbon dioxide from human respiration, which was calculated from the following expression (Luo et al., 2016; Sarbu and Sebarchievici, 2013) . Where M is the metabolic rate, also defined as the level of physical activity (met); r is the respiration coefficient (dimensionless) and is defined as the ratio between the volumetric rate at which CO2 is produced and the rate at which oxygen is consumed, its value depends mainly on the diet (0.83-1.0) and AD is the DuBois surface area (m 2 ) calculated from the height (H) in meters and body mass (W) in kg as follows: CO2 generation rate per person depends on age, gender, weight and metabolic activity. For teachers with a light and low activity level and an age range of 40 to 60 years, a value of 0.0062 l/s was selected, while for students with a light activity level and an age range of 16 to 30 years, a value of 0.0057 l/s was selected (Persily and de Jonge, 2017) . The analytical solution of equation (2) is shown below, where Cin is the initial concentration of the tracer gas. In order to solve equation (5), the steady state method was chosen to be used Haverinen-Shaughnessy et al., 2011; Zhong et al., 2019) . This method is based on the assumption that steady-state and well-mixed conditions are achieved. Furthermore, it assumes that the CO2 generation rate (that is, the occupants' number and activity) is constant over a sufficiently long period of time to reach the indoor equilibrium concentration. It also assumes that outdoor CO2 concentration and ventilation rate remain constant during measurement period (Batterman, 2017; Bekö et al., 2016) . ASTM (ASTM, 2018) suggests that the measured indoor equilibrium concentration Css should reflect at least 95% of the equilibrium value (that is, it is reached after 3 air changes). Using the mass balance equation, the steady-state concentration can be expressed as (Luther et al., 2018) : J o u r n a l P r e -p r o o f Where Css is the steady-state concentration (ppm), N is the number of persons present and and Gp is the average CO2 generation rate of a person (ml/s). Considering that all the tests performed lasted longer than 45 minutes and that during at least that time the number of students in the tests remained constant, the conditions for reaching steady state are fulfilled. In addition, in all cases the number of air renewals, as will be calculated later, was well above 3 (Hänninen et al., 2017; Persily, 2018) . Css was determined as the maximum average concentration over 5 minutes within the study window (Batterman, 2017) . Cex was calculated as the mean value of the outdoor CO2 concentration measurements taken for 5 minutes before and after each of the tests analysed. As could be seen, the mean difference between the CO2 values before and after the tests was not more than 4%, so this assumption was considered acceptable. Air exchange rate in the classroom was obtained by using the following expression: Once the real ventilation rate per test was obtained, the outdoor air flow rate per person was calculated. Real outdoor air flow rate, VR (l/s per person), was calculated as: Objective CO2 is defined as the predictable CO2 concentration based on enclosure volume and occupancy if a given number of ACHs of clean air from the enclosure were to be performed (Allen et al., 2020) . If the measured CO2 concentration is similar to the steady-state concentration, the ventilation objective is satisfied. If the CO2 concentration is higher than the steady-state concentration, the air exchange target is not reached and ventilation conditions should be revised. Given the variations in concentrations over the measured period, it is reasonable to assume a 20% deviation from the objective value before taking drastic actions (Minguillón et al., 2020) . The target CO2 value allows the establishment of a limit reference for ventilation control, as it indicates the CO2 level that must not be exceeded in order to guarantee an adequate level of ventilation, as proposed by Ilyas et. al. (Ilyas et al., 2015) . The calculation of the steady-state target CO2 concentration was carried out considering the values of the outdoor CO2 concentration (Cex) and CO2 generation rate per person described in the previous section. Steady-state objective CO2 J o u r n a l P r e -p r o o f (ppm) concentration was obtained using the following formula, where ACHobjective is the number of ACH required to ensure adequate ventilation. In this work, steady-state objetive CO2 concentration was calculated considering two situations, in the first case, the values of this parameter were determined for 5 ACH and in the second, the values for reaching an outdoor air flow of 14 l/s per person. In this second case, it was previously necessary to determine the number of renovations required for this outdoor air supply per person, applying Expression (7) (Allen et al., 2020; ASHRAE, 2019; Minguillón et al., 2020) . Thermal comfort is the state in which people consider themselves to be satisfied with their environment and can be assessed by a quantitative analysis through the heat balance model. It is therefore related to the overall heat balance of people and depends on physical activity and clothing, as well as air temperature, average radiant temperature, air velocity and relative humidity (AENOR, 2006) . PMV and PPD indicators are associated with heat balance model. Former represents the predicted mean vote on the wind chill scale of a group of people exposed to a certain environment (Fig. 4) . This implies considering as dissatisfied people those who voted cold (-3), cool (-2), warm (+2) or hot (+3) (AENOR, 2006; Fanger, 1970) . PPD establishes a quantitative prediction of the percentage of people who will not feel ambient satisfaction by noticing hot or cold sensations (Fanger, 1970) . This was quantitatively calculated using Expression (10). Thermal comfort conditions during the tests could be affected due to the ventilation strategies adopted and the low temperatures outside during the time of the year under study, with temperatures below 10˚C inside the classrooms, so it is recommended to assess the risk of cold stress by calculating the IREQ (required clothing insulation) index. IREQ index is defined at two levels of physiological overload. On the one hand, IREQ minimum defines the thermal insulation required J o u r n a l P r e -p r o o f to maintain the thermal equilibrium of the body at a lower than normal average body temperature level and, on the other hand, IREQ neutral defines the insulation for a thermal equilibrium temperature level (AENOR, 2009). For both PMV and IREQ calculations, the clothing of the classroom occupants was considered to correspond to an insulation of 1.56 clo and their metabolic activity level was assessed as sedentary activity with a value of 69.78 W/m 2 . Temporal evolution of CO2 concentration levels and air temperature as a function of the number of students present in classrooms at each moment was analyzed for a sample of the most representative exams carried out according to selected ventilation strategy. In addition, percentage dissatisfied (PPD) variation was collected as a function of the existing environmental conditions and it was found how the reduction in occupancy over time affected the concentration of carbon dioxide. The relationship between the different environmental and ventilation parameters was also assessed for each tests, taking into account the classrooms in which exams were held and the average outdoor temperature. Table 4 shows the results obtained for relative humidity, temperature and air velocity before, during and after the performance of all the tests studied. Pre-and post-test measurements were taken outside in the immediate vicinity of the school. To the work carried out by by Kephalopoulos et al., comfortable indoor classroom temperatures should be maintained, as far as possible, between 20 and 26˚C, regardless of the season and outside air temperature (Kephalopoulos et al., 2014) . In this study, average indoor temperatures ranged from 6 to 21˚C. As can be seen, only three of the Taking into account the method of assessing the concentration of CO2 levels, none of the average values exceeded 700 ppm, so ventilation during the tests was sufficient. In most cases 600 ppm was not exceeded, so the air quality was very good, in line with IDA category 1, except in one case. Maximum CO2 concentration value (808 ppm) was measured during test 5, however, the tolerable limit of 950 ppm for IDA category 2 was not exceeded (AENOR, 2008) . This point value was due to the ventilation strategy followed during this test, as detailed in section 3.5 of this work. Villanueva et al. found average CO2 concentrations of 539 ppm in pre-school classrooms, 565 ppm in primary classrooms and 661 ppm in secondary classrooms. These values were very similar to those obtained in this work (Villanueva et al., 2021) . However, in the case of the research carried out in schools in Switzerland and Latvia, CO2 concentration was much higher, with average values of around 2,000 ppm, and consequently air renewal in both cases was insufficient (Vassella et al., 2021; Zemitis et al., 2021) . Comparing CO2 concentrations, it can be verified that ventilation conditions in the present study are significantly better than those reported in most of the existing literature. In fact, CO2 levels recorded in this research were, on average, 3 times lower than those recorded in these European classrooms. Moreover, CO2 concentrations were found to be much lower, even compared to concentrations measured in classrooms located in other regions of Spain (Fernández-Agüera et al., 2019; Krawczyk et al., 2016) or in countries with similar climatic conditions, such as Portugal (Madureira et al., 2016) . As indicated above, calculation of steady-state CO2 concentration provides an indication of the levels that should not be exceeded depending on classroom volume and the set air renewal rate and depending on minimum air supply per person. As could be seen, only in test number 8, the measured mean CO2 concentration exceeded the steady-state objective for 5 ACH. Nevertheless, with the results for this parameter calculated for an input of 14 l/s per person, it could be seen that the situation was not problematic in any case, in fact, the margin in all cases was quite wide, this was proven by the results corresponding to the estimated external flow, which exceeds the 14 l/s per person by far and the associated stable CO2 levels, well above those measured in almost all cases. This showed that the recommendation to ensure at least 5 ACH may not be a good indicator when the occupancy of a classroom is low, as in the case of test 8, with adequate ventilation conditions and an actual number of renovations of 2.89. J o u r n a l P r e -p r o o f This can be explained by the fact that ACH does not depend on the number of enclosure occupants and should only be used for occupancies in the average range. ACH values required for an input of 14 l/s per person calculated were less than 5 in all tests, with values of up to 0.74 ACH in tests 11 and 16, the tests with the lowest occupancy ratio (0.01 persons per m 3 ). These values contrast with real ACH values that were much higher in almost all tests. Exam number 4 had the highest number of real ACH (19.36 renewals) and the maximum outdoor airflow per person (104.11 l/s per person). As could be seen, the lowest temperatures were also recorded in this test, compromising the thermal comfort of the occupants. All other ACH values and outdoor air flow per person were higher than those established to try to avoid COVID-19 infections inside the classrooms (Allen et al., 2020) . Along these lines, in other studies carried out for higher education buildings, between 3.7 and 39.8 outdoor air renewals were obtained depending on different ventilation configurations (de la Hoz-Torres et al., 2021). Table 6 presents corresponding results for the assessment of thermal conditions during the exams held in January. As inferred from the results, more than half of the tests were carried out in comfortable conditions and the rest in slightly cooler conditions. Only one of the tests (exam 4) was carried out in conditions outside the discomfort zone. In this case, it was assessed whether the situation during the test could be qualified as cold stress. IREQmin and IREQneu calculation values were 2.3 clo and 2.6 clo, respectively, and a recommended minimum exposure duration of 1.1 h and 1.5 h, respectively, which implies that with the test duration and clothing insulation considered, unsuitable situations could result (AENOR, 2009) . Measured values indicated that PPD in tests was just over 17%, values that are not considerably higher than those considered in the comfort range (limited by a 10% dissatisfaction rate). In tests with discomfort, PPD was estimated to be between 25-35%. Average PPD in exam 4, the worst conditions, was just over 70%, although there were times when virtually all test takers were dissatisfied with the thermal environment conditions. In studies carried out by other authors in the same pandemic and climate context as this work, they showed that more than 60% of the hours analyzed there were situations of thermal discomfort, with PPD ranging between 20% and 70% (Alonso et al., 2021) . J o u r n a l P r e -p r o o f and took measurements in winter in three regions of China. The most similar climatic conditions to those studied in this study are in the Shaanxi area, where 38.1% of the students rated the environment as neutral, and 31.9% rated it as slightly cool, so the percentage of votes of 1 and 0 was 70%, while 16.9% of the students rated the environment as cool or cold. Among the 177 male students, the average wind chill was 0.42, and among the 168 female students it was 0.48 (Wang et al., 2017) . Exam 14 was conducted following the general strategy of opening doors and windows in the classroom to maximise ventilation. It took place in the morning and lasted 2.25 hours. The number of students remained constant for just over an hour and then gradually decreased, with only two students remaining in class for the last half hour. In relation to CO2 concentration evolution in the classroom, a continuous increase of this parameter could be observed, which approximately stabilized one hour after the beginning of students' entry into the classroom. In the case of tests 17, 6 and 5 CO2 concentration increase was faster, this could be due to the higher number of students present in these tests compared to test 14. Maximum CO2 values did not reach 600 ppm and the average was 540.46 ppm so it can be considered as a situation with optimal IAQ conditions, it could also be observed that there was a decrease in the concentration as the students left the classroom (AENOR, 2008) . Another relevant factor to consider is the ACH produced, which was 5.6 with an outdoor air intake value of 51.74 l/s per person, both values above the recommendations to limit the possibilities of COVID-19 contagion (Allen et al., 2020; Minguillón et al., 2020) . Other studies using this same ventilation strategy have had between 7.4 and 9.4 outdoor air renewals for a 500 m 3 classroom and an occupancy of 48 people, 0.1 people per m 3 . (de la Hoz-Torres et al., 2021) . In relation to thermal environment, test was carried out with PPD values below 10% so J o u r n a l P r e -p r o o f conditions were comfortable (AENOR, 2006; Fanger, 1970) , with classroom temperature being close to 20˚C throughout the duration. Exam number 6 was conducted following the same strategy as the previous one, but in this case there was an intermediate break with all students leaving the classroom for a period of about 20 minutes. The students remained constant in number until half an hour before the break when they were phased out and then all returned to the classroom. In this case students' exit after the end of the exam was less progressive than in the case of exam 14. In relation to the evolution of CO2 concentration in the classroom, it could be seen that this is marked by the evolution of students' number inside the room. There were two rises associated with students entering the classroom at beginning of the exam and after the break, with maximum values approaching 650 ppm and an average of 588.17 ppm. It could also be seen that, although the students leaving the classroom before the break and rest time allowed a reduction in CO2 levels below 500 ppm, the values rose very quickly back up to pre-break concentration values. This showed that the effect of this stop was not significant on the IAQ during the test. The same effect was also found in the work of Zemitis et al. (Zemitis et al., 2021) . In relation to air renewals, these were 9.01 with a clean air supply per occupant of 42.39 l/s per person (values well above the minimum recommendations to guarantee adequate air quality and minimise COVID-19 infections) (Allen et al., 2020; Minguillón et al., 2020) . Exams 6 and 14 were held in classrooms E and B respectively, with identical dimensions, but with a slight variation in the configuration of doors and windows. This factor, in addition to the fact that the number of students in exam 6 were more than double compared to exam 14, could have influenced the differences in the CO2 concentrations measured. In relation to the thermal environment, according to the Fanger method the conditions were uncomfortable, slightly cold, with an average dissatisfaction rate of 35.58%, so it can be concluded that this test was carried out in relatively uncomfortable conditions (AENOR, 2006; Fanger, 1970) . This test was carried out in outside temperatures of a few degrees above zero, which meant that the conditions inside the classroom, with all the windows and doors open, were not suitable. Exam number 17 was held in room C, a room with the same characteristics as room E (where exams 5 and 6 were held) but located on the ground floor instead of the first floor. It lasted 3 hours and had 33 students. In this case, ventilation conditions were modified and only one of the existing windows was opened. CO2 concentration evolution followed a similar profile as in test 6 but in the case of test 17, CO2 levels reached were higher, with maximum concentrations of more than 700 ppm and average values of 624.51 ppm, values that were still within the quality parameters for air classified as IDA 1 and within the recommendations to avoid COVID-19 infections (AENOR, 2008) . Considering the ACH, 6.16, and the clean air supply per person, 20.46 l/s, also these criteria indicated that those conditions (although tighter than in the previous J o u r n a l P r e -p r o o f review) remained within recommendations (Allen et al., 2020; Minguillón et al., 2020) . In relation to the thermal environment, it was not problematic, average PPD of 5.09% (AENOR, 2006; Fanger, 1970) . In this case, in addition to the different ventilation strategy, the outside temperature, which averaged about 12˚C, played a decisive role. Finally, exam number 5, which, like exam number 6, took place in room E, lasted 2.40 hours and was taken by a total of 31 students, more than the 21 students who took exam number 6. In this case, windows were opened and closed according to the CO2 levels measured (windows were opened when CO2 values approached 650 ppm). CO2 concentration evolution followed a profile of rises and falls associated to the moments with and without window opening, with a maximum peak at the beginning of the test slightly exceeding 800 ppm (value corresponding to IDA 1) and successive lower peaks between 650 and 700 ppm, with the average concentration during the test being 606.99 ppm, all values above those recorded in tests 6 and 14. In this case, the air renovations were 11.56 and the outdoor air supply was 39.56 l/s per person, conditions that can, as in the other tests described above, be considered adequate (Allen et al., 2020; Minguillón et al., 2020 (Vassella et al., 2021) . In relation to thermal comfort, the opening and closing of the classroom windows led to similar discomfort values in tests 5 and 6, with outside temperatures between 4 and 5˚C lower. In any case, although reduction of ventilation improved indoor thermal conditions, they remained uncomfortable (AENOR, 2006; Fanger, 1970) . After describing and comparing the evolution of different parameters for four tests performed, the results were compared considering the average values measured in the different classrooms. Fig. 6 shows the mean values measured for PPD and the air changes for the tests carried out in each of the classrooms analyzed. Also shown in the graph are two lines marking the recommended ACH levels (AENOR, 2008; Minguillón et al., 2020) and PPD levels to be in the comfort zone (AENOR, 2006; Fanger, 1970) . The graph shows, as already indicated, that more than half of the classrooms had a dissatisfied percentage of less than 10%, indicating that they were conducted in comfortable conditions, with an ACH higher than the 5 recommended by the RITE (except in one case) (European Parliament, 2002; RITE, 2007) . On the other hand, although not decisive, an increase in discomfort was observed in many cases when the ACH was increased, in cases where a cold outside temperature was combined with a high number of ACH, which produced the worst situations of discomfort (tests 4, 5 and 6). In particular, test 4, the worst test for comfort conditions, was the one with lowest combined outdoor temperatures and highest ACH values. However, in those situations where the outside temperature was higher, the renovations did not significantly affect discomfort. J o u r n a l P r e -p r o o f were similar at quite different occupancy levels. In relation to classroom E, it was observed that similar temperature levels were achieved in tests 5 and 6 with much lower outside temperatures in the first case. This may have been due to the higher occupancy level, but in this case it was considered that it also affected the window opening and closing strategy during the test. Permanent closure strategy of part of the windows applied in tests 8 and 17 showed that the effect of this strategy, as mentioned above, had much less effect at higher outdoor temperature levels. 9 shows the values for the parameters maximum CO2 concentration, mean indoor temperature and PPD. Data were grouped according to whether the average external temperature was up to 6˚C, between 6 and 12˚C or above 12˚C. From analysis of data it was found that at outdoor temperatures above 12˚C in all cases the PPD indicated comfortable conditions (AENOR, 2006; Fanger, 1970) . CO2 concentration levels were in all these cases very close to 500 ppm, with the highest CO2 concentration values occurring in test 17, where a strategy of closing part of the windows was applied, which in this case was not considered necessary. Similarly, temperatures below 6˚C have always caused discomfort conditions, in this case, strategies such as closing part of the windows or opening and closing them may be of interest, as they improve the environmental conditions even if comfort is not achieved. In the case of very low temperatures, as was the case in test 4, excessive ventilation was not justified as it led to very cold conditions without the required air quality or outside air supply. In the range between 6 and 12˚C the comfort/discomfort conditions were determined by other factors such as occupancy, as in the case of test 3, which improved thermal environment conditions but worsened air quality. The study focused on the analysis of the adequacy of ventilation in accordance with the general air quality recommendations for educational centers and those specified to limit coronavirus infections and how they influenced thermal comfort conditions during examinations in winter periods. This limits its extension to other situations where windows do not have to be permanently open or other seasons. It would be interesting to extend the study to other periods. The building analysed has a specific location and characteristics, so the results could not be fully extended to other university centers on the campus. It would be interesting to extend the study to other buildings in order to compare results. In relation to the methods used, each class was assumed to be a separate zone with well mixed indoor air and could be characterised by a single measurement, this should be confirmed. The stationary method was chosen and a constant value for the CO2 concentration of the inlet air was considered, it would be interesting to perform the analysis using other methods such as decay or transient mass balance and continuous measurement of the outside CO2. The use of occupant-generated CO2 as a tracer gas has many advantages but can lead to untested measurement errors. The use of injected tracer gases could be of interest. In relation to the assessment of thermal comfort, standard-based methods were used and it would be interesting to assess the actual perceptions of the occupants by means of questionnaires. No account was taken of how factors relating to differences in the shape and configuration of openings in the classrooms analysed might affect them. Finally, this study focused on the assessment of ventilation and thermal comfort, it could be interesting to extend it also to the measurement of particulate matter. J o u r n a l P r e -p r o o f Like many other work spaces, on-site university teaching has been seriously threatened and modified in times of COVID-19. In this scenario, there are doubts about the appropriateness of returning to classroom-based teaching and assessment models. The uncertainty that a return to normality may cause justifies this study, which objectively analyses the risk of on-site assessment in times of pandemic. After the analyses performed, it can be concluded that the different ventilation strategies were correct in terms of CO2 concentration in all tests performed, with average CO2 concentration levels of between 450 and 670 ppm. In no case was the limit value set for category IDA 2 buildings, corresponding to educational establishments, exceeded, and in almost all cases an optimal IAQ corresponding to category IDA 1 was achieved. The maximum CO2 concentration value recorded in one of the tests was 808 ppm. However, these measures affected the thermal comfort of the occupants when outdoor environmental conditions were more unfavourable. In most cases, the number of real ACH above the recommended 5 for adequate ventilation was given. On the other hand, if the outdoor air flow per person is taken into account, in all cases values higher than the minimum established to try to avoid COVID-19 infection inside the classrooms, set at 14 l/s per person, were calculated. In this sense, it can be affirmed that the totally presential evaluation, in the terms in which it was carried out, does not put at risk the safety and health of students and teachers. In other words, in relation to the ACH value, it was considered that it is not a representative parameter for assessing IAQ when classroom occupancy was low. On the other hand, the occupancy ratio of all tests did not compromise the conditions set for a correct IAQ and had a slight influence on thermal comfort. Analysing in more detail the differences that occurred between different tests analysed, when outside temperature levels were above 12°C all tests were carried out in acceptable conditions of comfort irrespective of outside air supply or classroom occupancy, with a dissatisfaction rate of less than 10%. A significant influence on thermal comfort was observed for air changes when temperatures were below 6ºC, where a dissatisfaction rate of between 25 and 72% was observed. This influence was not noticeable at temperatures above 12°C. At temperatures below 12°C it is recommended to establish a ventilation strategy with opening and closing of windows, or to limit the number of open windows. This strategy should be complemented by the installation of CO2 meters (preferably with measurement of concentration values) to manage the strategy. At temperatures above 12°C, it is recommended to choose the complete opening of glazed openings if the risk situation is high, as thermal comfort is not significantly compromised. With low risks of infection, strategies of gradual opening and closing of windows can be chosen to avoid excessive energy losses. J o u r n a l P r e -p r o o f Therefore, in moderate or hot climates, there is no high cost, in terms of comfort, to carry out on-site tests with security guarantees. However, it is recommended that these on-site tests be limited to times or time slots with very low outside temperatures, considering alternatives such as holding these assessment tests in afternoon hours (normally with milder temperatures) or avoiding them on the coldest days of calendar. In seasons other than cold, the infection risk is not increased by using on-site evaluation while respecting distance and ventilation protocols, without compromising thermal comfort of the students. Ultimately, bearing in mind that ventilation protocols are still active in many countries, it is recommended that a strategy of generally opening windows when outside temperatures are mild should be considered in successive exams. The authors declare no conflict of interest relating to the material presented in this article its contents, including any opinions and/or conclusions expressed, are solely those of the authors. Informe sobre la borrasca Filomena y la ola de frío posterior Parte 1: Parámetros del ambiente interior a considerar para el diseño y la evaluación de la eficiencia energética de edificios incluyendo la calidad del aire interior Determinación e interpretación del estrés debido al frío empleando el aislamiento requerido de la ropa (IREQ) y los efectos del enfriamiento local UNE-EN 13779 Determinación analítica e interpretación del bienestar térmico mediante el cálculo de los índices PMV y PPD y los criterios de bienestar térmico local Ergonomía de los ambientes térmicos. Instrumentos de medida de las magnitudes físicas 2020. 5-step guide to checking ventilation rates in classrooms Recognizing and controlling airborne transmission of SARS-CoV-2 in indoor environments Indoor environmental quality of classrooms in Southern European climate Effects of the covid-19 pandemic on indoor air quality and thermal comfort of primary schools in winter in a mediterranean climate ANSI/ASHRAE Standard 62.1-2019: Ventilation for Acceptable Indoor Air Quality Indoor temperature, relative humidity and CO2 monitoring and air exchange rates simulation utilizing system dynamics tools for naturally ventilated classrooms ASTM D6245-18 Standard Guide for Using Indoor Carbon Dioxide Concentrations to Evaluate Indoor Air Quality and Ventilation Natural ventilation for infection control in health-care settings Indoor air pollution, physical and comfort parameters related to schoolchildren's health: Data from the European SINPHONIE study Review and Extension of CO2-Based Methods to Determine Ventilation Rates with Application to School Classrooms Ventilation rates in recently constructed U.S. school classrooms Diurnal and seasonal variation in air exchange rates and interzonal airflows measured by active and passive tracer gas in homes Use of food quotients to predict respiratory quotients for the doublylabelled water method of measuring energy expenditure Enfermedad por nuevo coronavirus, COVID-19. Situación actual How COVID-19 Spreads Multiple steady states in stack ventilation Analysis of Impact of Natural Ventilation Strategies in Ventilation Rates and Indoor Environmental Acoustics Using Sensor Measurement Data in Educational Buildings /91/EC of the European Parliament and of the Council of 16 December 2002 on the energy performance of buildings Thermal Comfort: analysis and applications in environmental engineering A field indoor air measurement of SARS-CoV-2 in the patient rooms of the largest hospital in Iran CO2 concentration and occupants' symptoms in naturally ventilated schools in mediterranean climate Control of CO2 in a naturally ventilated classroom School-based exposure to hazardous air pollutants and grade point average: A multi-level study Analysis of CO2 monitoring data demonstrates poor ventilation rates in Albanian schools during the cold season Association between substandard classroom ventilation rates and students' academic achievement Thermal comfort models and perception of users in free-running school buildings of East-Mediterranean region Influence of a CO2 Feedback System on Occupant Behavior in a Naturally Ventilated Space NTP 1064: Calidad del aire interior. Contaminantes biológicos (I): estrategia de muestreo NTP 741: Ventilación general por dilución El control de la ventilación mediante gases trazadores La ventilación como medida preventiva frente al coronavirus SARS-CoV-2 Informe sobre la situación de COVID-19 en España n o 1 Informe sobre la situación de COVID-19 en España n o 9 A field study of adaptive thermal comfort in primary and secondary school classrooms during winter season in Northwest China Institute for Health and Consumer Protection., European Commission. Directorate General for Health & Consumers., Regional Environmental Centre for Central and Eastern Europe Airborne bioaerosols and their impact on human health CO2 concentration in naturally ventilated classrooms located in different climates-Measurements and simulations Aerosol Measurement: Principles, Techniques, an Applications A case study on thermal comfort analysis of school building Effect of geometric parameters on ventilation performance in a dry room Revisiting an overlooked parameter in thermal comfort studies, the metabolic rate Investigating CO2 concentration and occupancy in school classrooms at different stages in their life cycle Indoor air quality in Portuguese schools: levels and sources of pollutants Transmission of SARS-CoV-2 by inhalation of respiratory aerosol in the Skagit Valley Chorale superspreading event Evaluación del riesgo de la transmisión de SARS-CoV-2 mediante aerosoles Carbon dioxide generation rates for building occupants Development of an Indoor Carbon Dioxide Metric Examination of thermal comfort in a hospital using PMV-PPD model Review of tracer gas-based methods for the characterization of natural ventilation performance: Comparative analysis of their accuracy Aspects of indoor environmental quality assessment in buildings Natural Ventilation Level Assessment in a School Building by CO2 Concentration Measures Aerosol and Surface Stability of SARS-CoV-2 as Compared with SARS-CoV-1 From spontaneous to strategic natural window ventilation: Improving indoor air quality in Swiss schools Assessment of CO2 and aerosol (PM2.5, PM10, UFP) concentrations during the reopening of schools in the COVID-19 pandemic: The case of a metropolitan area in Central-Southern Spain Roadmap to improve and ensure good indoor ventilation in the context of COVID-19 Methods for sampling and analysis of chemical pollutants in indoor air Supplementary publication to the screening tool for assessment of health risks from combined exposure to multiple chemicals in indoor air in public settings for children COVID-19 Safe return to school Trends in intake of energy and macronutrients in adults from Thermal comfort in primary school classrooms: A case study under subarctic climate area of Sweden Experimental investigation of thermal comfort and CO2 concentration in mosques: A case study in warm temperate climate of Yalova The Study of CO2 Concentration in A Classroom during the Covid-19 Safety Measures Indoor Environmental Quality Evaluation of Lecture Classrooms in an Institutional Building in a Cold Climate We would like to appreciate the Regional Government of Extremadura for the support toresearch groups (GR10151).