Mishra 2014_Volatile Organic Compounds-Characteristics, distribution and...

4                                                              N. Mishra et al. / Atmospheric Environment xxx (2014) 1e7


                                                                                              or close to one indicates significant outdoor source contribution
                                                                                              (Edwards et al., 2001). The geometric mean concentration of total
                                                                                              VOCs was used to calculate I/O ratios, which were found to be >1 in
                                                                                              the majority (i.e. 19) of the schools studied as can be seen in Fig. 3,
                                                                                              which implies that, in addition to outdoor sources, there were
                                                                                              dominant indoor sources in these schools.
                                                                                                  Only five schools (S02, S05, S08, S18 and S19) had I/O ratios <1,
                                                                                              which shows that outdoor VOC concentrations were higher than
                                                                                              indoor, and implying that outdoor sources had dominant contri-
                                                                                              bution. This can be explained by the combination of the important
                                                                                              factors: wind and traffic during school hour as well as the level of
                                                                                              indoor emissions. In S02, S05 and S08 the wind direction was
                                                                                              predominantly from the road to school and the number of vehicles
                                                                                              passing per hour was also higher i.e. 1418, 1096 and 1106, respec-
                                                                                              tively; this accounted for the high outdoor concentrations. S19 had
                                                                                              comparatively high traffic count (1609) though the wind direction
                                                                                              was from the school to the road whereas S18 had very low traffic
                                                                                              count (92 vehicles/hr) and the wind direction was also from the
     Fig. 3. Indoor/outdoor ratio (in increasing order) of total VOCs for each school.        school to the road. These two factors cannot explain the low value
                                                                                              of I/O ratio in these schools.
                                                                                                  It can also be seen from Fig. 4 that I/O ratio in seven schools (S16,
concentrations for indoor and outdoor total VOCs for all schools                              S23, S09, S17, S13, S12 and S11) was very high indicating dominant
were 125 mg/m3 and 71 mg/m3 respectively. The range for total VOCs                            indoor sources. There were evidences found inside the classroom
was (8e1070 mg/m3) indoors and (3e826 mg/m3) outdoors. In                                     during the measurements in these schools such as in S09 classroom
addition to this, individual concentrations of the most common                                was newly painted because the school was flooded, some renova-
compounds and their I/O ratios are presented in boxplots in                                   tion was done recently in S11. Similarly, there were much paper-
Supplementary Figure S5.                                                                      work activities done by the students hanging on the wall, windows,
   Fig. 2 presents a comparison between the total VOCs concen-                                over desk and benches in S12 and S13. There was a sink inside the
trations identified in this study and those reported in literature. It                         classroom in S17 where the use cleaning liquids, soaps was very
can be seen that the total VOC concentrations measured in this                                common. During the measurements in S16 and S23, cleaners were
study were either similar or higher than the studies performed in                             wiping desk, table using the spray. These local evidences support
other countries, but significantly lower indoors than those                                    for the high concentration indoors in these schools.
measured in South Korea. In the later study the samples were
collected from schools that had been recently renovated and                                   3.4. Source identification
reconstructed.
   In addition to the studies included in Fig. 3, Pegas et al. (2011a),                         Source identification of indoor and outdoor VOC concentrations
(2011b) reported the range for total VOCs (mg/m3) in different                                was carried out using the PCA method.
seasons in schools as (a) Indoor: 37e317 in Spring, 11e922 in
autumn, 84e2175 in winter (b) outdoor: 6e18 in spring, 5e50 in                                3.4.1. Outdoor VOCs
autumn, 7e22 in winter. Similarly, Sofuoglu et al., 2011 showed the                               As explained in the data analysis section, out of the 48 outdoor
comparison between outdoor playground and classrooms concen-                                  VOCs identified, the 16 species which were the most abundant
trations in their study (Classroom: 27 in fall, 111 in winter, 66 in                          compounds in the atmosphere were selected for further analysis.
spring; Outdoor playground: 22 in fall, 41 in winter, 24 in spring).                          The PCA anysis was carried out twice for outdoor data in order to
Although, season was not taken into consideration in our study, the                           see the changes made by including and excluding S03 but the re-
overall trend was similar to other studies showing that in all sea-                           sults were very similar therefore the results with S03 were
sons indoor levels were higher than outdoor.                                                  mentioned here. Five components/factors were extracted from the
   The I/O ratio is generally used to infer penetration of pollutants                         application of PCA to the data. The rotated component matrix for
into the indoor environment through doors and windows as well as                              outdoor VOCs is presented in Table 1. Varimax rotation with Kaiser
the presence of indoor sources (Tang et al., 2005). A ratio less than                         Normalisation was applied to determine the appropriate number of




                                           Fig. 4. Source contribution of each source to the total VOCs concentrations indoor and outdoor.



    Please cite this article in press as: Mishra, N., et al., Volatile Organic Compounds: Characteristics, distribution and sources in urban schools,
    Atmospheric Environment (2014), http://dx.doi.org/10.1016/j.atmosenv.2014.10.052