Probabilistic Prediction Models and Influence Factors of Indoor Formaldehyde and VOC Levels in Newly Renovated Houses
, Yaw-Shyan Tsay 3,†and Ching-Chang Lee 1,5,*,†
Round 1
Reviewer 1 Report
In the atmosphere-1675132 authors presented the results of a field study in which measurements of target VOCs in residences, were measured. Results were used to perform a risk assessment by using a probabilistic approach.
Overall, the manuscript should be improved in terms of with clarity of exposure and exhaustiveness. The introduction and results sections need to be consolidated. Some methodological doubts have been raised in the following comments, which I ask the authors to clarify. This could have a not negligible impact on the results obtained.
- Line 40: I imagine the construction of new buildings, it can be relevant, right?
- Line 44-46. Please consider that other than those reported here, previous studies on IAQ have considered some key pollutants or have sought to define prioritization criteria for the indoor monitoring of a wider number of air pollutants. Overall chemical pollutants considered as being of relevance for IAQ assessment in are substatially well known. See for example:
*Sérafin, Indoor Air. 31 (2021) 646–659
*Logue, et al. Indoor Air. 21 (2011) 92–109.
*Du, Wang. Aerosol Air Qual. Res. 20 (2020) 1851–1855.
*WHO - 2021, WHO global air quality guidelines: particulate matter (PM2.5 and PM10), ozone, nitrogen dioxide, sulfur dioxide and
carbon monoxide. https://apps.who.int/iris/handle/10665/345329
*Mandin et al. Sci. Total Environ. 579 (2017) 169–178
*WHO - Regional Office for Europe, 2020. https://apps.who.int/iris/handle/10665/334389
- Line67-64: Please consider to cite a reference of broader (international) validity and applicability (as WHO guidelines, just to cite an example) as well as the reference cited, valid but very local in scope
- Line 67-68: Please check ad update reference format (The main emission 67
source of formaldehyde is urea-formaldehyde resins (Hines et al., 1993) --> [NN]) - Line 72: "...the strongest correlation (Lee, 2004). Wu (2008) revealed..." Please check carefully the reference format through the manuscript
- Line 74-77: IMO some studies are available, albeit relatively scarce. At least some studies could be cited here.
- 2.3. Exposure and health risk assessment. I have serious doubts about this paragraph and methods. I believe more recent indications suggest that Using EPA's current methodology, it is unnecessary to calculate an inhaled dose when using dose-response factors in a risk assessment. and that This updated methodology recommends that risk assessors use the concentration of the contaminant in air (Cair) as the exposure metric (eg, mg / m3) instead of the intake of a contaminant in air based on inhalation rate and body weight (dose; eg, mg / kg-day).
However, inhalation risk assessments may require that an adjusted air concentration be used to represent continuous exposure.
See https://www.epa.gov/expobox/exposure-assessment-tools-routes-inhalation for further details. In addition, it can be estimated that often a calculation of HQ and HI is carried out considering the concentration in the air (and not the inhaled one), referring to guide values ​​of environmental concentration. Furthermore, why not also make a calculation of the carcinogenic risk? - Due to the doubts expressed in the previous point, the results and conclusions may require substantial changes, if the methods will be adapted to what is recommended. The approach used to present the results however is quite convincing and quite effective. I only suggest working on clarity of exposure and optimization of contents
Author Response
RESPONSES TO REVIEWERS
Ms. Ref. No.: atmosphere-1675132
We are grateful for the opportunity to revise our work. We have taken the comments on board to improve the manuscript. Please find below a point-by-point response to all comments.
Reviewer 1
In the atmosphere-1675132 authors presented the results of a field study in which measurements of target VOCs in residences, were measured. Results were used to perform a risk assessment by using a probabilistic approach. Overall, the manuscript should be improved in terms of with clarity of exposure and exhaustiveness. The introduction and results sections need to be consolidated. Some methodological doubts have been raised in the following comments, which I ask the authors to clarify. This could have a not negligible impact on the results obtained.
Line 40: I imagine the construction of new buildings, it can be relevant, right?
Response: 11 of new buildings processed to renovate were investigated in this study.
Line 44-46. Please consider that other than those reported here, previous studies on IAQ have considered some key pollutants or have sought to define prioritization criteria for the indoor monitoring of a wider number of air pollutants. Overall chemical pollutants considered as being of relevance for IAQ assessment in are substatially well known.
See for example:
*Sérafin, Indoor Air. 31 (2021) 646–659
*Logue, et al. Indoor Air. 21 (2011) 92–109.
*Du, Wang. Aerosol Air Qual. Res. 20 (2020) 1851–1855.
*WHO - 2021, WHO global air quality guidelines: particulate matter (PM2.5 and PM10), ozone, nitrogen dioxide, sulfur dioxide and carbon monoxide. https://apps.who.int/iris/handle/10665/345329
*Mandin et al. Sci. Total Environ. 579 (2017) 169–178
*WHO - Regional Office for Europe, 2020. https://apps.who.int/iris/handle/10665/334389
Response: Thank you for your comment. The sentence has been revised as “Among VOCs, benzene, ethylbenzene, toluene, styrene, and xylenes (BTX) have been iden-tified as major indoor air pollutants and have been included in the priority list of indoor compounds”. The suggested references have also been cited.
Line64-67: Please consider to cite a reference of broader (international) validity and applicability (as WHO guidelines, just to cite an example) as well as the reference cited, valid but very local in scope
Response: Thank you for your suggestion. The references have also been cited.
Line 67-68: Please check ad update reference format (The main emission source of formaldehyde is urea-formaldehyde resins (Hines et al., 1993) --> [NN])
Response: Thank you for your comment. The citation have been revised and included in the reference list.
Line 72: "...the strongest correlation (Lee, 2004). Wu (2008) revealed..." Please check carefully the reference format through the manuscript
Response: Thank you for your comment. The citation have been revised and included in the reference list.
Line 74-77: some studies are available, albeit relatively scarce. At least some studies could be cited here.
Response: Thank you for your comment. The citation have been added and included in the reference list [Lines: 74-78].
2.3. Exposure and health risk assessment. I have serious doubts about this paragraph and methods. I believe more recent indications suggest that Using EPA's current methodology, it is unnecessary to calculate an inhaled dose when using dose-response factors in a risk assessment. and that This updated methodology recommends that risk assessors use the concentration of the contaminant in air (Cair) as the exposure metric (eg, mg / m3) instead of the intake of a contaminant in air based on inhalation rate and body weight (dose; eg, mg / kg-day). However, inhalation risk assessments may require that an adjusted air concentration be used to represent continuous exposure. See https://www.epa.gov/expobox/exposure-assessment-tools-routes-inhalation for further details. In addition, it can be estimated that often a calculation of HQ and HI is carried out considering the concentration in the air (and not the inhaled one), referring to guide values ​​of environmental concentration. Furthermore, why not also make a calculation of the carcinogenic risk? Due to the doubts expressed in the previous point, the results and conclusions may require substantial changes, if the methods will be adapted to what is recommended.
Response: Thank you for your comment. In this study, we estimated the average daily doses for adult and children of exposure to formaldehyde and seven VOCs through the inhalation route. According to the Guidelines for Human Exposure Assessment (U.S. EPA, 2019), to estimate human exposure to contaminants in ambient or indoor air, information is needed about the exposed population(s), exposure pathways, and the concentrations of contaminants in air. When determining inhalation risk, inhalation rate and body weight might not be needed. The Superfund Program’s updated approach for determining inhalation risk eliminates use of inhalation rates when evaluating exposure to air contaminants. However, estimating the dose associated with intake from inhalation exposure is complicated because of the complex nature of the respiratory system as a portal of entry (U.S. EPA, 2009). Characteristics of average daily dose and risk of the exposed populations are important to define because they determine inhalation rate. In addition, inhalation rates will vary between children and adults due to differences in size, physiology, behavior, and activity levels. Under the exposure scenarios, the assessors could choose inhalation rate data that best represent the population for which exposures are being used to calculate the average daily potential dose from inhalation of a contaminant in air. The carcinogenic risks for adult and children have been included in Table 4.
References:
US EPA (United States Environmental Protection Agency). Risk Assessment Guidance for Superfund. Volume I: Human Health Evaluation Manual (Part F, Supplemental Guidance for Inhalation Risk Assessment). Final. (EPA/540/R-070/002 Publication 9285.7-82). 2009, Washington, D.C.: United States. Available online: https://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=P1002 UOM.TXT (accessed on April 9, 2022).
US EPA (United States Environmental Protection Agency). Guidelines for
Human Exposure Assessment. (EPA/100/B-19/001). 2019, Washington, D.C. United States. Available online: https://www.epa.gov/risk/guidelines-human-exposure-assessment (accessed on April 9, 2022).
The approach used to present the results however is quite convincing and quite effective. I only suggest working on clarity of exposure and optimization of contents
Response: Thank you for your suggestions. The manuscript has been revised seriously according your constructive comments.
Reviewer 2 Report
See the attached file.
Comments for author File: 
Comments.pdf
Author Response
RESPONSES TO REVIEWERS
Ms. Ref. No.: atmosphere-1675132
We are grateful for the opportunity to revise our work. We have taken the comments on board to improve the manuscript. Please find below a point-by-point response to all comments.
Reviewer 2
Atmosphere 1675132 Probabilistic Prediction Models and Influence Factors of Indoor Formaldehyde and VOCs Level in Newly Renovated Houses. The manuscript is an interesting and potentially useful assessment of formaldehyde and VOC concentrations in indoor air, within residential spaces. The work combines field sampling, regression analysis of factors, and inclusion of health and cancer risk assessments based on the measured concentrations. While the manuscript is generally good and a useful contribution to the field, there are a few significant areas where it requires improvement, especially in the areas of statistical analysis and prior literature. These, together with some minor comments, are listed below.
- Overall, the manuscript could use some additional editing for English grammar. Although it is generally good, some further editing would make it more perfect.
Response: Thank you for your comment. The certificate for language editing has been enclosed.
- There seems to be a mixture of citation styles for references, and these should be fixed.
Response: Thank you for your comment. The format of the reference citation has been checked across the manuscript according the instruction from Atmosphere.
- Line 98: the calculation of loading factors could use more explanation as to the methodology to permit reproduction by others.
Response: Thank you for your comment. The loading factor was divided the area of individual decoration material by total volume of each interior room [Lines: 101-103].
- Line 132: it seems unusual to use retention time to identify the VOC compounds when a GC/MS is used and could confirm the VOC identities by the mass spectra. Some further explanation might be recommended.
Response: Thank you for your comment. Both mass spectra and retention time were used to identified and quantified. The sentence has been revised as “The compounds of interest were identified and quantified by their retention time and mass spectra of calibration VOC standards (COMPRESSED GAS, N.O.S UN 1956, 10 ppm, procured from Linde) under specified chromatographic conditions” [Lines: 137-138].
- Line 179: more details on the multiple regression analysis methods used would be helpful. It seems that all predictor variables were left in the model. Why not use a type of stepwise regression to focus on the predictors with the best and significant impact? From the regression coefficients shown in the results sections (Table 3), it appears that many predictors have no statistical significance, yet they remain in the models shown in the supplementary materials.
Response: Thank you for your comment. In MLR model, the input variables should not be correlated with each other (multi-collinearity). All predictor variables were force entered into the models for predicting formaldehyde and specific VOCs and evaluated simultaneously due to those nonmodifiable variables including healthy green building materials, temperature control, and open window would imply the control strategies to reduce the potential health risks of formaldehyde and specific VOCs while renovated the newly house.
- Regression analysis: were the raw values of the predictor variables used (for example temperature in degrees Celsius), or were they rescaled in some manner?
Response: Thank you for your comment. In multiple linear regression, the raw values of all predictor variables were used.
- Table 1: the range of temperature and relative humidity measured for the samples is very narrow. It is difficult to see how any subsequent model could make use of these as predictor variables, as the small differences between high and low temperature (for example) are likely overwhelmed by analytical variability in formaldehyde and VOC measurements. This is possibly seen in Table 3 where temperature is not a significant predictor in the regression models (except for toluene). A stepwise regression may have likely eliminated this variable.
Response: Thank you for your comment. All the measurements were conducted in summer. The ranges of temperature and relative humidity measured for the indoor samples were narrow. The emission of the investigated indoor pollutants from decoration materials was reported that highly correlated to the local temperature [Lines: 313-316]. The limitation has been included in Discussion [Lines: 366-371]
- Line 196: this paragraph includes a repeat of some numbers that are already given in Table 2, which is not necessary.
Response: Thank you for your comment. The repeated numbers such as range have been removed.
- Table 2: the living rooms and bedrooms show a sample size of 9 and 10, respectively. However, line 95 indicates that 20 samples were collected (and Table 1), and line 186 mentions 10 living rooms and 12 bedrooms. Why are there differences in these quoted sample sizes?
Response: Thank you for your comment. The correct number of indoor air samples for or formaldehyde (n=20 and 20) and specific VOCs (n=16 and 20) from the living room and bedroom collected under opened and closed windows have been checked and modified [Lines: 98-99, 204; and Table 1 and Table 2].
- Table 3: as noted above, there are 10 predictors but only 5 or less are statistically significant for various measured VOCs. In some cases, only 2 of the predictors are significant. The adjusted R2 values are less than 50% in 4 of the 8 cited VOCs, indicating that the regressions are unable to account for the majority of the variance. Some further discussion or justification for this modelling seems required, and a discussion of how it compares with other attempts to model these difficult situations would be helpful.
Response: Thank you for your comment. The present study represents the preliminary development of the risk-based probabilistic predicting model to characterize quantitatively the concentrations of formaldehyde and specific VOCs contributed by the decoration materials in the newly renovated houses. In addition, the established models effectively implied to control strategies for the influence factors of formaldehyde and specific VOCs for reduce health risk in the newly renovated houses. The discussion and limitation regarding the risk-based probabilistic predicting model has been mentioned [Lines: 309-326, 359-371].
- Table 3: the coefficients shown in Table 3 don’t seem to correspond to any values given in the prediction models listed in the supplementary materials in Table S3.
Response: Thank you for your comment. The values which round off to the 3rd decimal place given in the prediction models (Table S3) were corrected according to the coefficients reported in Table 3.
- Line 271: this paragraph is difficult to read and it repeats many of the numbers that are already shown in Table 4. It would be better to just focus on discussion of the main items to be noted from Table 4.
Response: Thank you for your comment. The paragraph have been revised and discussed focus on the findings from Table 4 [Lines: 255-276].
- Figure 1: the interpretation of this figure is not clear or easy to understand. It seems to show a positive effect of an “Open window” on cancer risk, for example, similar to the presence of a wooden floor. Is it not expected that an open window decreases cancer risk? Perhaps the terminology is not well explained.
Response: Thank you for your comment. “Open window” showed the negative effects on non-cancer risk (-1.5% to -9.1% for hepatic effect; -0.2% to -1.7% for urinary effect; -7.7% to -11.1% for developmental effect) and cancer risk (-8.9% to -14.1%) in the results of sensitive analyses. Figure 1 has been revised.
- Line 313: it is not quite clear that this is indeed the “first study to model the relationship between health risk and formaldehyde…”. In fact, there are a number of recent publications related to this general area that the manuscript does not cite. It would be recommended that the manuscript consider these also, and in the discussion compare them with the current results in a bit more comprehensive way. It can then be made clearer where the similarities, differences, and novelty exists in the current work. A few examples of recent literature on formaldehyde that were not cited are given below, and others on VOCs should also be sought and included.
Li, Baizhan, et al. "An investigation of formaldehyde concentration in residences and the development of a model for the prediction of its emission rates." Building and environment 147 (2019): 540-550.
Bourdin, Delphine, et al. "Formaldehyde emission behavior of building materials: on-site measurements and modeling approach to predict indoor air pollution." Journal of hazardous materials 280 (2014): 164-173.
He, Zhangcan, et al. "An improved mechanism-based model for predicting the long-term formaldehyde emissions from composite wood products with exposed edges and seams." Environment international 132 (2019): 105086.
Zhang, Zi-Feng, et al. "Indoor occurrence and health risk of formaldehyde, toluene, xylene and total volatile organic compounds derived from an extensive monitoring campaign in Harbin, a megacity of China." Chemosphere 250 (2020): 126324.
Liang, Weihui. "Long-term indoor formaldehyde variations and health risk assessment in Chinese urban residences following renovation." Building and Environment 206 (2021): 108402.
Huang, Shaodan, et al. "Indoor formaldehyde concentrations in urban China: preliminary study of some important influencing factors." Science of the Total Environment 590 (2017): 394-405.
Response: Thank you for your comment. The sentence has been revised as “The probabilistic and sensitive predicting models for formaldehyde and selected VOCs were established effectively to reduce health risk in the newly renovated houses.” [Lines: 373-374]. The suggested references have been included in Discussion and reference list.
Round 2
Reviewer 1 Report
The authors have considered and responded to all the reviewers' comments. As far as I'm concerned, the changes introduced in the revised manuscript are satisfactory. I have no other comments
Author Response
RESPONSES TO REVIEWERS
Ms. Ref. No.: atmosphere-1675132
Reviewer 1
The authors have considered and responded to all the reviewers' comments. As far as I'm concerned, the changes introduced in the revised manuscript are satisfactory. I have no other comments
Response: Thank you for the agreement of publication in Atmosphere.
Reviewer 2 Report
The authors have made significant improvements to the manuscript and clarified many of the uncertainties. I have only a few minor suggestions:
- Probably the term "MCS P95" should be defined in the Abstract.
- Line 95 "residents" should be "residences"
- Line 317 "agree" should be "agreement"
Author Response
RESPONSES TO REVIEWERS
Ms. Ref. No.: atmosphere-1675132
Reviewer 2
The authors have made significant improvements to the manuscript and clarified many of the uncertainties. I have only a few minor suggestions:
Response: Thank you for the opportunity to revise our work. We have taken the comments on board to improve the manuscript. Please find below a point-by-point response to all comments.
- Probably the term "MCS P95" should be defined in the Abstract.
Response: Thank you for your comment. The term “MCS P95”has been defined in Abstract [Line 28].
- Line 95 "residents" should be "residences"
Response: Thank you for your comment. The word “residents has been revised as “residences” [Lines: 96].
- Line 317 "agree" should be "agreement"
Response: Thank you for your comment. The word “agree” has been revised as “agreement” [Lines: 320].
