Building and Environment 186 (2020) 107317

Contents lists available at ScienceDirect

Building and Environment

journal homepage: http://www.elsevier.com/locate/buildenv

Carbon dioxide generation rates of different age and gender under various
activity levels
Liu Yang a, b, Xueni Wang a, b, Minghui Li a, b, Xiang Zhou c, Shichao Liu d, Hui Zhang e,
Edward Arens e, Yongchao Zhai a, b, *
a
State Key Laboratory of Green Building in Western China, Xi’an University of Architecture and Technology, Xi’an, Shanxi, 710055, China
b
College of Architecture, Xi’an University of Architecture and Technology, Xi’an, Shaanxi, 710055, China
c
School of Mechanical Engineering, Tongji University, Shanghai, 200092, China
d
Department of Civil and Environmental Engineering, Worcester Polytechnic Institute, Worcester, MA, 01609, USA
e
Center for the Built Environment, UC Berkeley, Berkeley CA, 94720, USA

A R T I C L E I N F O A B S T R A C T

Keywords: Indoor carbon dioxide (CO2) is a critical parameter in the design and control of ventilation, and in monitoring
CO2 generation rates fresh air levels in buildings. Building occupants are primary sources for indoors CO2, and the rate at which CO2 is
Age and gender differences generated depends on the occupants’ physical activities. In the past, CO2 generation rates have been indirectly
Metabolic rate
calculated from metabolic rate data. There has been little systematic experimental determination of human CO2
Indoor air quality
Ventilation standards
generation rates at different activity levels. This study experimentally determines human CO2 generation rates
for use in building design and control applications. CO2 generation rates were obtained by a high accuracy in­
direct calorimetry system on 99 human subjects aged from 20 to 70, who were asked to perform 18 activities
(including office, household, and walking activities) in a climatic chamber controlled at 26 ◦ C. CO2 generation
rates increase significantly as activity levels increase. Age was found to have a small effect, while gender dif­
ferences were significant. Women produced less CO2 than men at the same activity levels, mainly attributable to
their lower body masses. Regression models were developed based on activity levels and age for each gender to
predict CO2 generation rates. Significant discrepancies were found between measured CO2 generation rates
predicted by the models in ASHRAE 62.1–16 and ASTM D6245-18, especially at higher activity levels. These
findings provide fundamental data for designers and researchers in ventilation and indoor air quality field, and
future ventilation standard revisions.

1. Introduction
Workers and students spend large numbers of their time at work or
school. Poor indoor air quality (IAQ), often caused by insufficient
ventilation, adversely affects their perception of air quality, productivity
or performance, and health [1–3]. In designing and controlling space
ventilation systems, indoor CO2 concentration has often been used as a
surrogate for the ventilation rate per occupant [4–7]. Indeed, minimum
ventilation rates were originally specified to eliminate pollutants
generated by occupants [8–11]. In current standards, the required
ventilation rate could be calculated for most buildings by a steady-state
mass-balance equation using CO2 concentrations and CO2 generation
rate [12,13].
Occupants are the primary source of CO2 in occupied indoor spaces
[9]. CO2 generation rate is typically derived from metabolic rate using
equations given in the ASHRAE Handbook [14]. However, such methods
have not been experimentally validated for different ages and genders.
Recently, Persily and de Jonge [15] presented a comprehensive table of
CO2 generation rates for different age and gender groups, based on basal
metabolic rates (BMR). It also included the effects of body size, body
composition, diet, and level of physical activities. However, this
approach is also an indirect method relying on metabolic rate data [16,
17] and assumed values of respiratory quotient (RQ) lacking support of
experimental data.
To date, there are very limited experimental data on CO2 generation
rates from building occupants. Lawrence and Braun [18] presented a
methodology for estimating occupancy schedules to calculate CO2
source generation rates but did not associate it with activity levels. Qi

* Corresponding author. State Key Laboratory of Green Building in Western China, Xi’an University of Architecture and Technology, Xi’an, Shanxi, 710055, China.
E-mail address: zhaiyongchao@xauat.edu.cn (Y. Zhai).

https://doi.org/10.1016/j.buildenv.2020.107317
Received 19 July 2020; Received in revised form 16 September 2020; Accepted 22 September 2020
Available online 25 September 2020
0360-1323/© 2020 Published by Elsevier Ltd.
L. Yang et al. Building and Environment 186 (2020) 107317

et al. [19] carried out CO2 generation rate experiments in an air-tight
chamber with 44 young Chinese people under two levels of activity,
sitting quietly and standing relaxed. The study found that the CO2
generation rates by real occupants have a factor of 0.75 for Chinese
women and of 0.85 for Chinese men over the predictions using the
ASHRAE method. A recent study by Zhai et al. [20] found that dis­
crepancies between measured metabolic rates and those given in the
ASHRAE Standard were the primary causes of inaccuracies in the CO2
generation rate calculations using in the ASHRAE equations, or when
using the new method proposed by Persily and de Jonge [15].
More experimental data are needed on the CO2 generation rates
using realistic building occupants of different ages and genders, engaged
in typical office and household activities, and in walking. The current
research responds address the identified limitations in the literature by
examining a broad range of experimental subjects and tasks using a high
accuracy wearable device capable of directly measuring the concentra­
tion and flow rate of respired CO2.
2. Methods
2.2. Facilities and measurements
2.2.1. Climate chamber
All tests were performed in a climate chamber (Fig. 1a) to stan­
dardize the thermal environments. The chamber (4.5 m × 3.9 m × 2.7
m) simulates an office environment. It controls temperature to an ac­
curacy of ±0.2 ◦ C and RH ± 5%. The air and radiant temperatures in the
chamber was maintained steadily at 26 ◦ C during all tests to minimize
the effect of temperature on CO2 generation rates. The relative humidity
was controlled at 50%, and air speed less than 0.1 m/s. The indoor CO2
concentration of the chamber was between 400 and 600 ppm when two
research assistants and one subject were present during each test.
The air-conditioning system of the chamber was a dedicated all air
system. After the fresh air is conditioned in temperature and humidity by
the air-conditioning unit, it is supplied into the room via perforated
plates in the ceiling and returned via a raised floor with perforated plates
too. This design ensures that the indoor thermal environment within the
chamber is evenly distributed. We used 100% fresh air in this study with
an air change rate 60 times per hour.

2.1. Human subjects 2.2.2. CO2 measurements

The human subject testings were conducted from April to June 2019
in the climate-controlled chamber at Xi’an University of Architecture
and Technology. Ninety-nine healthy volunteers (50 women and 49
men, aged 20–70 years) participated in the experiment, with nearly ten
male or female subjects in each age category (Table 1). Most of the
subjects were university faculty members, staff, or students. Participa­
tion was voluntary, and the subjects were provided with written consent
before the experiments. The subjects were allowed to quit the tests if
they felt uncomfortable at any time during the test. Because the tests
involved some higher metabolic rates, two research assistants were in
the chamber during the tests to ensure the safety of the subjects.
Subjects’ anthropometric data are summarized in Table 1. Subjects’
height (measured with SECA 213, accuracy 1 cm) and weight (measured
with PESA CB 2.2–100, PESA Ltd, Beijing, China, accuracy 0.002 kg)
were obtained before each test. Their body fat content was measured by
a body fat scale (measured with OMRON V-body HBF-701-SH with 0.1%
accuracy). Body mass index (BMI) and body surface area (ADu) were
calculated based on the measured height and weight.
Women of all age groups were significantly lower in both height and
weight than men, but they had about 5–10% higher body fat than men.
Regarding BMI, the youngest group (20–29 yrs) had about 4% lower BMI
and body fat than the older groups. There were no differences in ADu
among all age groups for both genders.
A high accuracy wearable metabolic system (COSMED K5, COSMED
S.r.l., Italy) was used for measuring VCO2 directly. (Fig. 1b). The
wearable mask (Fig. 1c) guarantees maximum seal with almost no air
leaks. The one-way inspiratory valve makes sure that all inhalations are
ambient room air, and all exhalations are collected to analyze the CO2
changes. The CO2 sensor in COSMED K5 is a rapid-response non-
dispersive infrared sensor (NDIR), a spectroscopic sensor often used as a
gas detector. It has an accuracy of <3% at 1–7 L/min. A micro-dynamic
mixing chamber (IntelliMET™, US Patent 9581539) was used to mea­
sures gas exchange with a micro-dynamic mixing chamber mode that
was applied to this experiment.
Before each test, the K5 was first warmed up for 15 min, after which
gas sensors were calibrated using both ambient air (21% O2, 0.03% CO2,
and 78% N2) and a standard gas (5% O2, 16% CO2 and 79% N2), and the
flowmeter was calibrated using a standard 3-Liter calibration syringe.
Subjects put on the face mask that is connected to the K5 instrument.
The headgear was adjusted to comfortable tightness levels requested by
subjects. The sealing property of the mask was examined by the
researchers.
2.2.3. Activities tested
Eighteen physical activities were divided into three types (Table 2):
sitting and standing office work, household activities, and five walking
speeds on a treadmill. The chamber was divided into three parts for
these three types of activities and a small area for a subject to lie down to

Table 1
Subjects anthropometric information.
Age Group Sample size Age (yr) Height (m) Weight (kg) ADu (m2) BMI Body fat Fat-free mass (kg)
(kg/m2) (%)

Female
20–29 10 24.7 (1.5) 1.61 (0.04)**** 52.2 (3.5)*** 1.5 (0.1)**** 20.0 (1.3) 25.4 (2.9)** 38.9 (2.1)****
30–39 10 32.2 (2.2) 1.63 (0.02)** 56.3 (7.1)* 1.6 (0.1)* 21.3 (2.5) 27.6 (3.4)* 40.6 (4.0)***
40–49 10 46.4 (2.4) 1.57 (0.05)*** 60.7 (8.0)** 1.6 (0.1)*** 24.6 (3.0) 31.3 (3.0)* 41.5 (4.0)****
50–59 10 54.6 (2.0) 1.59 (0.04)** 63.0 (9.25) 1.6 (0.1)* 24.8 (3.8) 34.0 (4.3)** 41.0 (4.0)****
60–69 10 62.7 (3.3) 1.55 (0.06)**** 59.6 (8.5) 1.6 (0.1)** 24.7 (2.7) 35.1 (3.6)*** 38.5 (4.6)****
Male
20–29 10 23.4 (2.4) 1.73 (0.04) 67.7 (8.2) 1.8 (0.1) 22.7 (2.8) 19.3 (4.2) 54.4 (5.3)
30–39 10 32.2 (3.1) 1.70 (0.05) 66.9 (12.6) 1.8 (0.2) 23.1 (3.3) 22.2 (5.3) 51.5 (7.2)
40–49 9 45.8 (2.6) 1.67 (0.06) 74.2 (11.0) 1.8 (0.1) 26.4 (3.4) 25.2 (4.4) 55.2 (6.0)
50–59 10 54.4 (3.0) 1.65 (0.06) 69.1 (7.0) 1.8 (0.1) 25.4 (2.7) 24.2 (3.8) 52.2 (3.6)
60–69 10 64.8 (2.6) 1.70 (0.04) 67.2 (9.4) 1.8 (0.1) 23.2 (3.1) 23.6 (7.9) 50.7 (4.6)

BMI (body mass index) = weight (kg)/height (m)2.

ADu (skin surface area) = 0.202 × (weight)0.425 × (height)0.725 [21].
All values are mean (SDs). * gender differences, p＜0.05；** gender differences, p＜0.01；*** significant gender differences, p＜0.001；**** significant gender
differences, p＜0.0001.

2
L. Yang et al. Building and Environment 186 (2020) 107317

Fig. 1. Experimental set-up and metabolic rate measurement equipment.

Table 2 Table 3
List of activities during the test. Experimental protocol.
Lying, Sitting, Standing, Household Walking Activity Description Duration
quiet office office activity (level)
Lying, quiet Lying flat in a dark and quiet environment 20 min
Lying Sitting, quiet Standing, Watching TV Walking, 2 Sitting, quiet Relaxing in a sitting position 8 min
down quiet km/h Sitting, typing Typing in a sitting position 8 min
Sitting, Standing, Using Walking, 3 Sitting, writing Writing in a sitting position 8 min
typing typing smartphone km/h Sitting, Reading a book in a sitting position 8 min
Sitting Standing, Vacuuming Walking, 4 reading
writing filing km/h Standing, Relaxing in a standing position in front of a 8 min
Sitting, Folding clothes Walking, 5 quiet height adjusting table
reading km/h Standing, Typing in a standing position in front of a 8 min
Window Walking, 6 typing height adjusting table
cleaning km/h Standing, Filing the same pile of files (1 kg) in the 8 min
filing standing position
Watching TV Watching TV of a random clip of a movie 8 min
stabilize his/her metabolism. Eight indoor office activities were tested in in sitting relaxing position
the office activity area: lying quietly, sitting quietly, reading sitting, Using Using a smartphone while sitting relaxed 8 min
smartphone
writing sitting, typing sitting, standing quietly, typing standing, filing Vacuuming 2
Vacuuming a fixed area (5.13 m ) at 8 min
standing. The household activity area was furnished for a total of five subjects’ self-selected pace
activities, watching TV, using smartphone, vacuuming, folding clothes, Folding Folding clothes into a specified style in 8 min
and cleaning a window. A treadmill was used for walking at five clothes subjects’ self-selected pace
Window Wiping a fixed area of window at subjects’ 8 min
different speeds (2–6 km/h, respectively). For the sitting condition,
cleaning self-selected pace
subjects were seated in front of a typical office desk in a plastic mesh Walking Walking on a treadmill at five speeds 8 min for each
chair that provided negligible additional insulation (Fig. 2b). For the consecutively: 2, 3, 4, 5, 6 km/h speed, no break
standing condition, subjects stood at a height-adjustable desk (IKEA
SKARSTA) (Fig. 2c). For the vacuuming activity, subjects used a hand­
held vacuum cleaner (Deerma DEM-DX600S, 2.7 kg) to clean the carpet moderate to vigorous physical activity the day before the test, as well as
in a fixed area. For the walking condition, subjects walked on a treadmill alcohol and caffeine-containing drinks. When the subject arrived at the
(LifeSpan TR1200B, LifeSpan Fitness/P. C. E. Inc. USA) at a set speed climate chamber, he/she changed into test clothes (0.6 clo, including a
and checked regularly by a tachometer (Fluke 931), again at a height- long-sleeved shirt, long pants, sneakers, and socks). Then he/she entered
adjustable desk (Fig. 2f). All activities that required typing on a the test chamber.
keyboard were not tested on the 60–70 age group because they were The subject first laid down on the top of a blanket on the floor for 20
generally less familiar with computer-based work. min. Then the face mask was attached and VCO2 measurements were
taken in the lying position for the last 8 min of the 20min. He/she was
then asked to perform sitting and standing office tasks in random order.
2.3. Experimental protocol After a short break, he/she was asked to perform the household tasks,
again in random order. Following this, he/she started the walking
The experimental protocol is shown in Table 3. Each subject conditions, in which the speed stepped up sequentially from 2 to 6 km/h.
participated in one 180 min experiment that took place between the Subjects over 60 were not required to complete all the walking
hours of 7:00 a.m. and 7:00 p.m. They were asked to refrain from

Fig. 2. Test activity conditions. a. Lying down, b. Sitting conditions, c. Standing conditions, d. Window cleaning, e. Folding clothes, f. Walking.

3
L. Yang et al.
conditions, to avoid possible health impacts.
2.4. Statistical methods
Statistical analysis was done using Graphpad Prism 8 for Windows
(GraphPad Software, San Diego, California US). Experimental data were
expressed as Mean (SD). Because it is apparent that CO2 generation rates
are positively associated with physical activity, and it is hard to interpret
the results of three-way ANOVA. Here we use two-way ANOVA with LSD
post hoc test was to test the main effects of gender and age, and their
interaction at each activity level. For comparisons with predicting
models in ASHRAE and ASTM, analysis of covariance (ANCOVA) was
used. Significance was accepted at P < 0.05.
3. Results and discussions
3.1. Mean values over time
Fig. 3 shows measured CO2 generation rates change over the 8-min
duration of each tested activity. Results from all subjects were aver­
aged in the plots. It is seen that in all cases, the measured CO2 generation
rates reached steady states after approximately 3 min. For the analysis
below, only the average steady-state values of the last 5 min of mea­
surements were used.
3.2. Mean steady-state values at different activities levels
Tables 4–6 present the steady-state CO2 generation rates (in Mean
and SD) at different activities for each gender and age group, arranged
by lying and office activities (Table 4), household (Table 5), and walking
(Table 6). CO2 generation rates were the lowest at rest (lying, quiet) for
both genders, increased slightly by body position change (sitting and
standing postures), then increased notably during walking, reaching the
highest values at 6 km/h. The office activities all had low generation
Fig. 3. Measured CO2 generation rates at all test conditions over time.
4
Building and Environment 186 (2020) 107317
rates, averaging 0.25 L/min. Among the household activities, the CO2
generation for standing and moving tasks (vacuuming, window clean­
ing, folding clothes) was nearly double that of the sitting tasks (watching
TV, using smartphone, and the office tasks). The average CO2 generation
was 0.41 L/min. For walking activities, there was a steady increase in
CO2 generation rates more than doubled as the walking speed increased
2 km/h to 6 km/h. The relative variation (SD/Mean) for each condition
shows that the degree of data dispersion does not vary according to the
intensity of the activity. The variations for all conditions ranged from
0.20 to 0.27 in the same gender/age groups, indicating that large inter-
individual differences among the subjects.
3.3. Effects of age and gender
Fig. A1 shows the distribution of CO2 generation rates, and Table A1
to A3 show the statistical analysis on age and gender effects for each
physical activity level. For all test conditions, there were no significant
effects of the interaction of age and gender. The effects of age and gender
are presented below.
3.3.1. Age
There were significant age effects in only 7 of the 18 activities tested,
including lying quiet, sitting writing, standing quiet, standing typing,
and walking at 2, 3, and 6 km/h. Post-hoc multiple comparisons show
that the differences are mainly found between 20 and 30 yrs group to the
40–50 and 50–60 yrs age group in the walking activities, with CO2
generation rates increase as age increases. However, the elder 60–70 yrs
group show similar CO2 generation rates as the 20–30 yrs group, which
is lower than the middle age groups. This may reflect the selection
criteria of subjects recruitment. In our study, participants who were
aged over 60 are most free of diseases and have exercise habits. Most
people over 60 years of age are retired and have more time to exercise,
unlike people in their 30s–50s who are of working age. Therefore, some
elder subjects over 60 may have better fitness levels and lower weight
L. Yang et al. Building and Environment 186 (2020) 107317

Table 4
Mean (SD) CO2 generation rates (L/min) at lying and office activity levels.
Age group Lying, quiet Sitting, quiet Sitting, typing Sitting, writing Sitting, reading Standing, quiet Standing, typing Standing, filing

Male
20–29 0.23 (0.06) 0.27 (0.07) 0.28 (0.05) 0.28 (0.05) 0.27 (0.07) 0.26 (0.07) 0.28 (0.05) 0.29 (0.05)
30–39 0.26 (0.06) 0.28 (0.07) 0.31 (0.08) 0.30 (0.07) 0.28 (0.07) 0.31 (0.08) 0.31 (0.07) 0.31 (0.06)
40–49 0.28 (0.06) 0.29 (0.07) 0.29 (0.05) 0.30 (0.04) 0.29 (0.05) 0.31 (0.07) 0.29 (0.05) 0.31 (0.05)
50–59 0.28 (0.06) 0.29 (0.07) 0.32 (0.07) 0.34 (0.06) 0.29 (0.06) 0.32 (0.06) 0.33 (0.06) 0.34 (0.05)
60–69 0.24 (0.06) 0.28 (0.07) 0.28 (0.05) 0.27 (0.08) 0.28 (0.06) 0.31 (0.07)
Female
20–29 0.19 (0.05) 0.21 (0.05) 0.23 (0.06) 0.23 (0.05) 0.22 (0.05) 0.22 (0.05) 0.22 (0.05) 0.24 (0.06)
30–39 0.23 (0.05) 0.23 (0.05) 0.23 (0.05) 0.25 (0.04) 0.24 (0.04) 0.25 (0.05) 0.27 (0.05) 0.25 (0.05)
40–49 0.20 (0.04) 0.22 (0.05) 0.22 (0.05) 0.23 (0.06) 0.22 (0.04) 0.23 (0.06) 0.21 (0.04) 0.25 (0.05)
50–59 0.20 (0.05) 0.22 (0.05) 0.23 (0.05) 0.25 (0.05) 0.23 (0.06) 0.23 (0.06) 0.23 (0.07) 0.25 (0.06)
60–69 0.19 (0.04) 0.20 (0.05) 0.22 (0.03) 0.21 (0.04) 0.22 (0.04) 0.24 (0.04)

than those in the 30–50 middle age group.

Table 5
Mean (SD) CO2 generation rates (L/min) at household activity levels.
3.3.2. Gender
Age Watching Using Vacuuming Window Folding For all 18 activities, there were significant gender differences, with
group TV smartphone cleaning clothes
women generate significantly less CO2 (around 20% less at office ac­
Male tivities and up to 35% less at walking activities) than men across all
20–29 0.25 0.26 (0.06) 0.47 (0.08) 0.50 0.41
activity conditions. The differences in body composition between men
(0.06) (0.10) (0.09)
30–39 0.26 0.27 (0.07) 0.49 (0.11) 0.46 0.43
and women might explain the observed gender difference [22–25].
(0.06) (0.10) (0.09) Women have significantly lower body mass than men across all age
40–49 0.26 0.28 (0.05) 0.53 (0.11) 0.54 0.44 groups (Table 1) therefore need less O2 for conducting the same physical
(0.06) (0.11) (0.08) activity and then lower CO2 generations.
50–59 0.28 0.31 (0.05) 0.50 (0.14) 0.52 0.44
(0.05) (0.13) (0.12)
60–69 0.25 0.26 (0.07) 0.46 (0.11) 0.44 0.38 3.3.3. Regression models
(0.06) (0.08) (0.08) According to the previous analysis, since gender differences were
Female significant and age also had a significant effect on some of the activity
20–29 0.20 0.22 (0.05) 0.35 (0.07) 0.33 0.32
levels. For ease of use, we selected activity level and age as independent
(0.05) (0.07) (0.08)
30–39 0.22 0.24 (0.05) 0.37 (0.08) 0.37 0.32
variables and used stepwise regression to create equations by gender
(0.05) (0.08) (0.09) (Equations (1) and (2)). It provides a simple method to obtain CO2
40–49 0.21 0.22 (0.05) 0.38 (0.08) 0.37 0.33 values based on age, gender, and metabolic rate. These equations are
(0.05) (0.07) (0.06) also intended for determining ventilation rates for spaces, estimating the
50–59 0.22 0.22 (0.05) 0.37 (0.09) 0.37 0.30
indoor CO2 levels, or guiding automated ventilation controls when oc­
(0.05) (0.09) (0.05)
60–69 0.19 0.21 (0.05) 0.40 (0.08) 0.37 0.32 cupancy information is available.
(0.04) (0.09) (0.05) ( )
Male:VCO2 = 0.264MET − 0.001Age − 0.061, R2 ⋅ = ⋅0.969 (1)
( )
Female:VCO2 = 0.230MET − 0.001Age − 0.046 ⋅⋅ R2 ⋅ = ⋅0.969 (2)
Table 6
Mean (SD) CO2 generation rates (L/min) at walking activity levels.
where VCO2, CO2 generation rate per person, L/min;
Age 2 km/h 3 km/h 4 km/h 5 km/h 6 km/h MET, metabolic rate for different activity levels, met;
group Age, 20–70, yrs.
Male
20–29 0.53 0.59 0.68 0.85 1.15
(0.09) (0.12) (0.14) (0.17) (0.23) 3.4. Comparisons with ASHRAE Std. 62.1–2106 and ASTM D6245-
30–39 0.59 0.64 0.76 0.92 1.25 2018
(0.11) (0.11) (0.14) (0.17) (0.23)
40–49 0.68 0.74 0.85 1.10 1.35
In practice, CO2 generation rates are often derived from the physical
(0.16) (0.16) (0.18) (0.27) (0.26)
50–59 0.68 0.73 0.84 1.03 1.30 activity levels according to the prediction equations in ASHRAE Stan­
(0.16) (0.16) (0.19) (0.22) (0.25) dard 62.1 [12] and ASTM D6245 [13]. ASHRAE Standard 62.1 shows
60–69 0.66 0.71 0.79 0.98 1.22 the relation between CO2 generation rate and metabolic rate, guiding
(0.17) (0.18) (0.17) (0.20) (0.31)
the calculation of the outdoor air flow rate in ventilation design. While
Female
20–29 0.41 0.47 0.55 0.70 0.95
the recently revised ASTM D6245-18 entirely referenced study [15] to
(0.07) (0.08) (0.07) (0.11) (0.16) predict CO2 generation rates. The prediction equations are listed in
30–39 0.42 0.49 0.59 0.76 1.03 Table 7, along with our empirical equations.
(0.08) (0.08) (0.12) (0.17) (0.22) When compared the CO2 predicting equation with the ASTM D6245
40–49 0.47 0.53 0.60 0.79 1.04
[13,15], we simplify the calculation by calculating each age group
(0.12) (0.12) (0.12) (0.20) (0.19)
50–59 0.53 0.56 0.63 0.76 1.08 separately in Equations (1) and (2), we bring in the age variable as in­
(0.13) (0.11) (0.11) (0.17) (0.29) termediate age (eg: 25 in age 20–29). CO2 generation rates of all the
60–69 0.54 0.59 0.67 0.86 1.18 prediction equations are expressed in L/min unit.
(0.10) (0.08) (0.10) (0.16) (0.23) Fig. 4 shows the rate of CO2 production by age, by gender, compared
with ASHRAE Std. 62.1 [12] and ASTM D6245-18 [13]. Compared to
the values given in the ASHRAE Standard 62.1, the CO2 generation rates
for men in our study are close to the values from the standard. For

5
L. Yang et al. Building and Environment 186 (2020) 107317

Table 7 activity levels, and RQ. While BMR may vary with age and gender [26,
Prediction equations of CO2 generation prediction rates. 27], it is doubtful that the same decline trends also apply for regular
CO2 generation prediction equations Input Parameters office and walking activities since many studies reported no age or
gender differences in metabolic rates at varies standardized activity
ASHRAE 62.1 VCO2 (L/min) = 0.258⋅MET Metabolic rate
[12] (met) levels [28–30]. Second, the metabolic rate table cited in ASTM
ASTM D6245 V (L/min) = RQ∙BMR∙MET 0.0341 RQ D6245-18 is different from the ones used in ASHRAE 55 or ISO 8996.
[13] BMR (MJ/day) The definition or 1 MET in Ref. [16] is 3.5 ml/min VO2 consumption,
Metabolic rate while it is 58 W/m2 in ASHRAE 55 and ISO 8996. It is suspicious that
(MET)
The current Male Metabolic rate
these metabolic data could be used interchangeably without any con­
study VCO2(L/min) = 0.264⋅MET - 0.001⋅Age - (met); version in units. Third, the approach if Ref. [15] assumes RQ as a con­
0.061 Age (yrs) stant 0.85. In our experiment, the RQ was not always 0.85 but changed
Female along with activity levels (Fig. 5).
VCO2(L/min) = 0.230⋅MET - 0.001⋅Age -
0.046
3.5. Potential applications in room ventilation designs

women, the CO2 generation rates from our study are lower than the
standard value for all age groups. According to the previous analysis, the
experimental CO2 generation differs by age and gender, and the ASH­
RAE standard method does not characterize CO2 generation rates for
different ages or gender. Therefore, the ASHRAE method may over­
estimate the room ventilation rate for spaces occupied by both genders.
Compared with ASTM D6245-18 [13], the results of our experiment
are higher for both men and women. The discrepancies are small at
lower activity levels but become highly evident at higher activity levels.
In Ref. [15], there is also a tendency for CO2 generation rates to decrease
with age in study [15], which is different from what we found in our
study. There may be three reasons for such differences. First, Ref. [15]
Tried to derive occupants’ CO2 generation rates from human BMR,
CO2 generated by building occupants is widely used in ventilation
design [31–34]. From the discussion above, it is obvious that CO2 gen­
eration rates in our tests are slightly lower than the values presented in
ASHRAE Standard 62.1 for men and significantly lower for women. In
this case, the outdoor air flow rate calculated by ASHRAE method would
be overestimated. The overestimation will result in energy wasted to
process excess outside air. The impact of the difference of CO2 genera­
tion tare on ventilation rate will be demonstrated quantitatively below.
Assuming there are the same numbers of men and women in the
space, and the mean age is 35. Then we insert Equations (1) and (2) into
equation D1 in Appendix D of ASHRAE Standard 62.1–2016 and arrive
at Equation (3) that shows the relation between outdoor fresh airflow
rate per person and activity levels (mean of a male and a female).

Fig. 4. Comparison of experimental data with ASHRAE standards 62.1 [12] and ASTM6245-18 [13] for different age groups.

6
L. Yang et al. Building and Environment 186 (2020) 107317

the outdoor airflow rate was overestimated and would lead to higher
energy consumption.
In addition to being used in airflow calculations, indoor CO2 con­
centration is an important indicator of indoor air quality [35–37]. The
CO2 generation rates provided in this experiment can be used to evaluate
indoor CO2 levels with the information on occupancy density and
ventilation rate.

3.6. Limitations

Despite the relatively large sample size (99 participants), only 20

subjects were tested (ten men and ten women) in each age group. Also,
this study only recruited adults in good health conditions. Therefore, our
results may only apply to such a population. For more specific appli­
cations, futures’ studies are needed that have larger sample sizes and
include adolescents and children with different ethnic backgrounds and
different levels of health conditions. In addition, our study was limited
to light to moderate activity levels. Higher-level metabolic activities
should also be tested in further studies.

Fig. 5. Respiratory Quotient (RQ) at different activity levels. 4. Conclusions

Human CO2 generation rates were directly measured at 18 stan­

0.5⋅(VCO2 )female + 0.5⋅(VCO2 )male
Vo = (3) dardized physical activity levels (including lying and office activity,
Cs − Co household activity, and walking from 2 km/h to 6 km/h) on 99 human
subjects aged 20 to 70 for both genders. We found:
where.
Vo, Outdoor airflow rate per person, L/min.
1. The measured CO2 generation rates are positively associated with
Cs, CO2 concentration in the space, L (CO2)/L (air).
physical activity levels. Age has a small effect on CO2 generation
Co, CO2 concentration in outdoor air, L (CO2)/L (air).
rates, while gender differences were significant. Women tend to
In the calculation, the permissible indoor CO2 concentrations are
produce less CO2 than men, and the differences are greater at higher
600 ppm, 800 ppm, 1000 ppm, 1500 ppm, 2000 ppm. The CO2 con­
activity levels.
centrations in outdoor air typically range from 300 to 500 ppm [12];
2. Deviations exist between the measured CO2 generation rates and the
therefore, 400 ppm is chosen for the calculation. Based on Equation (3),
values from the ASHRAE ventilation Standard 62.1 and ASTM
the outdoor airflow rates are calculated and compared with the corre­
D6245-18. ASHRAE prediction equations tend to overestimate actual
sponding values based on Standard 62.1 (Fig. 6). It can be seen that
CO2 generation rates, especially for women at higher activity levels.
when maintaining indoor CO2 concentration at 600–2000 ppm, the re­
ASTM D6245-18, on the other hand, tend to underestimate the actual
sults of Equation (3) are all lower than the ASHRAE Standard 62.1
CO2 generation rates for both men and women.
prediction, by 7.97–30.18%. At lower activity levels, the difference
3. Simple linear equations that predict CO2 generation rates from
between the calculated and standard values of outdoor airflow rate is
physical activity, age, and gender are developed for use in room
310–350 L/min, while at moderate activity levels, the difference is
ventilation rate estimations. The calculated minimum outdoor
350–575 L/min. The higher the activity level, the lower the calculated
airflow rates based on the model of this study are 8%–20% lower
value is compared to the standard value, which means in most buildings,
than those specified by ASHRAE Std. 62.1.
4. The findings in this experimental study provide fundamental data for
designers and researchers in ventilation and indoor air quality field,
and for future ventilation standard revisions.

Declaration of competing interest

The authors declare that they have no known competing financial

interests or personal relationships that could have appeared to influence
the work reported in this paper.

Acknowledgments

The work was supported by National Key R&D Program of China

(2018YFC0704500), National Natural Science Foundation of China
(51978553), and Key Research and Development Program of Shaanxi,
China (2018ZDCXL–SF–03-05).

Appendix

Fig. 6. Calculated minimum required outdoor airflow rate per person for
different indoor CO2 levels (600 ppm, 800 ppm, 1000 ppm, 1500 ppm, and
2000 ppm).

7
L. Yang et al. Building and Environment 186 (2020) 107317

Fig. A1. Box-plot of CO2 generation rates among different age groups of women and men for all physical activity levels.

Table A1
Two-way ANOVA statistics for lying and office activities. by gender and age

Sum of Square DF Mean Square F Statistic P value

Lying, quiet
Age 0.02 4 0.01 F (4, 88) = 2.88 P < 0.05
Gender 0.07 1 0.07 F (1, 88) = 36.30 P < 0.0001
Age × Gender 0.01 4 0.00 F (4, 88) = 0.96 P = 0.4358
Sitting, quiet
Age 0.01 4 0.00 F (4, 88) = 1.18 P = 0.3252
Gender 0.09 1 0.09 F (1, 88) = 57.52 P < 0.0001
Age × Gender 0.00 4 0.00 F (4, 88) = 0.58 P = 0.6746
Sitting, writing
Age 0.02 4 0.01 F (4, 89) = 2.81 P < 0.05
Gender 0.07 1 0.07 F (1, 89) = 40.81 P < 0.0001
Age × Gender 0.01 4 0.00 F (4, 89) = 1.20 P = 0.3150
Sitting, reading
Age 0.01 4 0.00 F (4, 87) = 1.87 P = 0.1235
Gender 0.06 1 0.06 F (1, 87) = 37.57 P < 0.0001
Age × Gender 0.00 4 0.00 F (4, 87) = 0.19 P = 0.9415
Sitting, typing
(continued on next page)

8
L. Yang et al. Building and Environment 186 (2020) 107317

Table A1 (continued )
Sum of Square DF Mean Square F Statistic P value

Age 0.01 3 0.00 F (3, 71) = 2.25 P = 0.0904

Gender 0.10 1 0.10 F (1, 71) = 49.52 P < 0.0001
Age × Gender 0.01 3 0.00 F (3, 71) = 0.90 P = 0.4448
Standing, quiet
Age 0.02 4 0.01 F (4, 88) = 2.68 P < 0.05
Gender 0.09 1 0.09 F (1, 88) = 43.68 P < 0.0001
Age × Gender 0.01 4 0.00 F (4, 88) = 0.69 P = 0.5980
Standing, typing
Age 0.03 3 0.01 F (3, 71) = 4.81 P < 0.05
Gender 0.08 1 0.08 F (1, 71) = 35.44 P < 0.0001
Age × Gender 0.01 3 0.00 F (3, 71) = 1.76 P = 0.1617
Standing, filing
Age 0.01 4 0.00 F (4, 88) = 1.55 P = 0.1941
Gender 0.11 1 0.11 F (1, 88) = 51.31 P < 0.0001
Age × Gender 0.00 4 0.00 F (4, 88) = 0.67 P = 0.6176

Table A2
Two-way ANOVA statistics for household activities by gender and age

Sum of Square DF Mean Square F Statistic P value

Watching TV
Age 0.01 4 0.00 F (4, 88) = 1.99 P = 0.1037
Gender 0.07 1 0.07 F (1, 88) = 40.42 P < 0.0001
Age × Gender 0.00 4 0.00 F (4, 88) = 0.47 P = 0.7592
Using smartphone
Age 0.01 4 0.00 F (4, 87) = 2.16 P = 0.0802
Gender 0.08 1 0.08 F (1, 87) = 53.46 P < 0.0001
Age × Gender 0.00 4 0.00 F (4, 87) = 1.23 P = 0.3050
Vacuuming
Age 0.02 4 0.01 F (4, 88) = 0.25 P = 0.9089
Gender 0.79 1 0.79 F (1, 88) = 38.03 P < 0.0001
Age × Gender 0.04 4 0.01 F (4, 88) = 0.48 P = 0.7517
Window cleaning
Age 0.03 4 0.01 F (4, 88) = 1.03 P = 0.3942
Gender 0.37 1 0.37 F (1, 88) = 48.29 P < 0.0001
Age × Gender 0.05 4 0.01 F (4, 88) = 1.64 P = 0.1721
Folding clothes
Age 0.01 4 0.00 F (4, 88) = 0.68 P = 0.6103
Gender 0.27 1 0.27 F (1, 88) = 54.77 P < 0.0001
Age × Gender 0.02 4 0.00 F (4, 88) = 0.91 P = 0.4595

Table A3
Two-way ANOVA statistics for walking activities by gender and age

Sum of Square DF Mean Square F Statistic P value

Walking, 2 km/h
Age 0.21 4 0.05 F (4, 90) = 3.21 P < 0.05
Gender 0.58 1 0.58 F (1, 90) = 34.68 P < 0.0001
Age × Gender 0.01 4 0.00 F (4, 90) = 0.15 P = 0.96
Walking, 3 km/h
Age 0.23 4 0.06 F (4, 88) = 3.95 P < 0.05
Gender 0.64 1 0.64 F (1, 88) = 44.59 P < 0.0001
Age × Gender 0.02 4 0.01 F (4, 88) = 0.39 P = 0.8127
Walking, 4 km/h
Age 0.24 4 0.06 F (4, 90) = 1.45 P = 0.2238
Gender 0.19 1 0.19 F (1, 90) = 4.62 P ¼ 0.0342
Age × Gender 0.05 4 0.01 F (4, 90) = 0.29 P = 0.8830
Walking, 5 km/h
Age 0.34 4 0.09 F (4, 90) = 0.99 P = 0.4187
Gender 1.06 1 1.06 F (1, 90) = 12.16 P ¼ 0.0008
Age × Gender 0.35 4 0.09 F (4, 90) = 1.00 P = 0.4081
Walking, 6 km/h
Age 0.73 4 0.18 F (4, 70) = 3.64 P < 0.05
Gender 1.09 1 1.09 F (1, 70) = 21.79 P < 0.0001
Age × Gender 0.12 4 0.03 F (4, 70) = 0.59 P = 0.6696

References
[1] World Health Organization, The right to healthy indoor air—report on a WHO
meeting, Bilthoven, NL, Europ. Health Targets 10 (2000) 13.
[2] M.G. Apte, Associations between indoor CO2 concentrations and sick building
syndrome symptoms in US office buildings: an analysis of the 1994-1996 BASE
study data, Indoor Air 10 (4) (2000).
[3] Z. Bakó-Biró, P. Wargocki, D. Wyon, P.O. Fanger, Poor indoor air quality slows
down metabolic rate of office workers, Proc. Indoor Air (2005) 76–80.

9
L. Yang et al.
[4] A. Persily, Challenges in developing ventilation and indoor air quality standards:
the story of ASHRAE Standard 62, Build. Environ. 91 (2015) 61–69.
[5] A.K. Persily, Evaluating Building IAQ and Ventilation with Indoor Carbon Dioxide,
American Society of Heating, Refrigerating and Air-Conditioning Engineers, 1997,
pp. 1–2505.
[6] S. Shriram, K. Ramamurthy, S. Ramakrishnan, Effect of occupant-induced indoor
CO2 concentration and bioeffluents on human physiology using a spirometric test,
Build. Environ. 149 (2019) 58–67.
[7] A. Kabirikopaei, J. Lau, Uncertainty analysis of various CO2-Based tracer-gas
methods for estimating seasonal ventilation rates in classrooms with different
mechanical systems, Build. Environ. 179 (2020), 107003.
[8] U. Satish, M.J. Mendell, K. Shekhar, T. Hotchi, D. Sullivan, S. Streufert, W.J. Fisk, Is
CO2 an indoor pollutant? Direct effects of low-to-moderate CO2 concentrations on
human decision-making performance, Environ. Health Perspect. 120 (12) (2012)
1671–1677.
[9] P. Fanger, Ventilation and body odor. Proc. Of an Engineering Foundation
Conference on Management of Atmospheres in Tightly Enclosed Spaces, ASHRAE,
1983.
[10] P. Smith, Determination of ventilation rates in occupied buildings from metabolic
CO2 concentrations and production rates, Build. Environ. 23 (2) (1988) 95–102.
[11] P. Kapalo, Ľ. Mečiarová, S. Vilčeková, E. Krídlová Burdová, F. Domnita, C. Bacotiu,
K.-E. Péterfi, Investigation of CO2 production depending on physical activity of
students, Int. J. Environ. Health Res. 29 (1) (2019) 31–44.
[12] ASHRAE, Standard 62.1-2016 Ventilation for Acceptable Indoor Air Quality,
American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc.,
Atlanta, GA, 2016.
[13] ASTM, Standard Guide for Using Indoor Carbon Dioxide Concentrations to
Evaluate Indoor Air Quality and Ventilation, American Society for Testing and
Materials, West Conshohocken, PA, 2018. D6245-18.
[14] ASHRAE Handbook Fundamental, American Society of Heating, Refrigerating and
Air-Conditioning Engineers, Inc., Atlanta, GA, USA, 2015.
[15] A. Persily, L. de Jonge, Carbon dioxide generation rates for building occupants,
Indoor Air 27 (5) (2017) 868–879.
[16] B. Ainsworth, W. Haskell, S. Herrmann, Compendium of physical activities: a
second update of codes and MET values, Med. Sci. 43 (8) (2011) 1575–1581.
[17] Food and Agriculture Organization of the United Nations (FAO), Human energy
requirements, report of a joint FAO/WHO/UNU expert consultation, Food Nutr.
Tech. Rep. Series 1 (2001). Geneva.
[18] T.M. Lawrence, J.E. Braun, A methodology for estimating occupant CO2 source
generation rates from measurements in small commercial buildings, Build.
Environ. 42 (2) (2007) 623–639.
[19] M.W. Qi, X.F. Li, L.B. Weschler, J. Sundell, CO2 generation rate in Chinese people,
Indoor Air 24 (6) (2014) 559–566.
10
Building and Environment 186 (2020) 107317
[20] Y. Zhai, M. Li, S. Gao, L. Yang, H. Zhang, E. Arens, Y. Gao, Indirect calorimetry on
the metabolic rate of sitting, standing and walking office activities, Build. Environ.
145 (2018) 77–84.
[21] D. DuBois, A formula to estimate the approximate surface area if height and body
mass be known, Arch. Intern. Med. 17 (1916) 863–871.
[22] P. Kapalo, F. Domnita, C. Bacotiu, M. Podolak, The influence of occupants’ body
mass on carbon dioxide mass flow rate inside a university classroom – case study,
Int. J. Environ. Health Res. 28 (4) (2018) 432–447.
[23] W.C. Adams, Influence of age, sex, and body weight on the energy expenditure of
bicycle riding, J. Appl. Physiol. 22 (3) (1967) 539–545.
[24] M.T. Korhonen, A. Mero, H. Suominen, Age-related differences in 100-m sprint
performance in male and female master runners, Med. Sci. Sports Exerc. 35 (8)
(2003) 1419–1428.
[25] A.S. Yue, J. Woo, K.W. Ip, C.M. Sum, T. Kwok, S.S. Hui, Effect of age and gender on
energy expenditure in common activities of daily living in a Chinese population,
Disabil. Rehabil. 29 (2) (2007) 91–96.
[26] J.A. Harris, F.G. Benedict, A biometric study of human basal metabolism, Proc.
Natl. Acad. Sci. U. S. A 4 (12) (1918) 370.
[27] W. Schofield, Predicting basal metabolic rate, new standards and review of
previous work, Human nutrition, Clin. Nutr. 39 (1985) 5–41.
[28] J. Durnin, L. Namyslowski, Individual variations in the energy expenditure of
standardized activities, J. Physiol. 143 (3) (1958) 573.
[29] J.V.G.A. Durnin, R. Passmore, Energy, Work and Leisure, Energy, Work and
Leisure, 1967.
[30] S. Strickland, S. Ulijaszek, Energetic cost of standard activities in Gurkha and
British soldiers, Ann. Hum. Biol. 17 (2) (1990) 133–144.
[31] I. Turiel, Occupant-generated CO2 as an Indicator of Ventilation Rate, 1980.
[32] A.K. Melikov, D.G. Markov, Validity of CO2 Based Ventilation Design, E3S Web of
Conferences, EDP Sciences, 2019.
[33] T. Lu, A. Knuutila, M. Viljanen, X. Lu, A novel methodology for estimating space air
change rates and occupant CO2 generation rates from measurements in
mechanically-ventilated buildings, Build. Environ. 45 (5) (2010) 1161–1172.
[34] O. Seppänen, W. Fisk, M. Mendell, Association of ventilation rates and CO2
concentrations with health andother responses in commercial and institutional
buildings, Indoor Air 9 (4) (1999) 226–252.
[35] C.A. Erdmann, K.C. Steiner, M.G. Apte, Indoor carbon dioxide concentrations and
sick building syndrome symptoms in the BASE study revisited: analyses of the 100
building dataset, Proceedings of Indoor Air 2002 (2002) 443–448.
[36] S.S. Korsavi, A. Montazami, D. Mumovic, Indoor Air Quality (IAQ) in Naturally-
Ventilated Primary Schools in the UK: Occupant-Related Factors, Building and
Environment, 2020, 106992.
[37] X. Zhang, P. Wargocki, Z. Lian, C. Thyregod, Effects of exposure to carbon dioxide
and bioeffluents on perceived air quality, self-assessed acute health symptoms, and
cognitive performance, Indoor Air 27 (1) (2017) 47–64.