Original Paper

Indoor and Built Environment

Indoor Built Environ 2008;17;6:543-551

Accepted: July 31,2008

Longitudinal Study of Indoor Particulate Matter and its Relationship to Outdoor Concentrations in New Delhi, India Sumeet Saksena a

R. Uma b

"The East-West Center, Honolulu, USA bThe Energy and Research Institute, New Delhi, India

Key Words indoor air· indoor-outdoor relation' particulate matter· office • New Delhi

longitudinal correlations between indoor and outdoor concentrations, the results of our study were contrary. The variation in outdoor concentration of RSP could explain only 25% of the variation in indoor concentra­ tion. Our results strongly support the belief that indoor­ outdoor relationships cannot be generalized, but are strongly dependent on factors such as type of building, ventilation, climate among others.

Abstract Current models used to estimate the health effects and burden of outdoor urban air pollution assume a strong day-to-day correlation between outdoor and indoor concentration of the pollutant(s) of interest. Studies of indoor-outdoor relationships, based on a day-to-day longitudinal design, in developing countries have been rare. Our study is a comparatively long longitudinal study conducted over three and a half years. Indoor levels of Total Suspended Particulates (TSP) and Respirable Suspended Particulates (RSP), dso = 10 Ilm, were measured in an office building and also just outside the building. The mean values of indoor TSP and RSP concentrations, though much higher than those observed in similar micro-environments in devel­ oped countries, were less than the outdoor concentra­ tions but higher than the Indian ambient standards. However, while studies in developed countries (and in some developing countries) have observed strong

Many epidemiological studies have documented asso­ ciations between particulate matter air pollution and several health effects including mortality, hospital admis­ sions, respiratory symptoms, and lung function [1-6]. Similar associations have even been documented in developing countries of Asia [7]. Most of these have been time series studies, which relate day-to-day variation in air pollution to day-to-day variation in health end­ points. Typically in these studies, exposure assessment is based on fixed site measurements in ambient air. It has been suggested that particulate matter concentrations

© SAGE Publications 2008 Los Angeles, London, New Delhi and Singapore DOl: 10.1177j1420326X08097294 Accessible online at http://ibe.sagepub.com

Dr Sumeet Saksena

The East-West Center, 1601 East West Road, Honolulu, HI, 96848, USA,

Tel. +1-808-9447249, Fax +1-808-9447298,

E-Mail [email protected]

Introduction

from fixed sites correlate poorly with personal exposures [8]. Some studies have found values of 0.06 [9] and 0.07 [10] for the correlation between personal and ambient Respirable Suspended Particulates (RSP), dso = 10 !-lm. In a large survey in the US the correlation between 24 h averaged personal and ambient PM lO was found to be 0.48 [11]. If the variation in outdoor levels of particulate matter is not strongly linked to variation in personal exposures, the use of outdoor as a surrogate for personal exposures would tend to misclassify personal exposures, leading to attenuation in exposure-response relations [12]. It is important to note that in most studies, the correlation between personal exposures and outdoor concentrations has been calculated cross-sectionally. Data are collected from a group of subjects by measuring personal exposures of different subsets of subjects on different days (leading to different ambient concentrations) and measuring each subject a limited number of times. Next, one correlation coefficient is calculated, using all measurements from all subjects and days. This correlation is influenced by the variation in personal exposure among subjects. Since time series studies relate day-to-day variations in outdoor concentrations to day-to-day variations in health end­ points, the correlation between personal and ambient concentrations within persons over time, is more relevant than the variation among persons [13]. This correlation may be better because some factors that can cause variation among subjects, such as smoking habits, are less variable within a subject over time, and therefore, they are mainly responsible for the variation across subjects. People spend most of their time indoors and it is well established. that indoor levels of pollutants may be different from outdoor levels. This raises the question of how well ambient concentrations reflect human exposure. The fact that past studies, mostly done in developed countries, seem to find good correlations between ambient concentrations and health effects indicates that at least the changes in total exposure seem to be adequately reflected by changes in ambient air pollution levels. The relation­ ship between outdoor and indoor concentrations is hard to generalize, however, because it depends on local characteristics, including building construction, ventilation, and existence of indoor sources. It is very likely that differences in climate, economic circumstances, building characteristics, etc. may result in quite different indoor-outdoor relations in developing countries. While some studies in India have surveyed the indoor air quality in homes [14-17] very few have studied the indoor air quality in offices.

544

Indoor Built Environ 2008;17:543-551

The aim of this study was to measure the correlation coefficient between indoor concentration of Total Sus­ pended Particulate matter (TSP) and RSP and outdoor concentrations measured at two locations: one just outside the building and the other at a well established central fixed site monitoring station. A secondary aim was to determine the seasonal patterns in the indoor-outdoor relationships.

Methods

The study was conducted in a relatively new office building (1994) in south Delhi's Lodi Estate. This office tower (Darbari Seth Block) is part of a complex of buildings called the India Habitat Centre (IHC) located on Lodi Road. The land-use pattern around the IHC is a mix of residential and commercial. Many international agen­ cies, cultural organizations, government offices (the CGO complex) and educational institutions are located within a 3 km radius. The IHC has offices, cultural centers and restaurants (not in the Darbari Seth Block though). Lodhi road is a four-lane east-west thoroughfare with a motorized vehicle count of 5333 vehicles per hour [18]. The Darbari Seth Block has six floors. The first four floors have only offices, but the other two floors have a mix of offices and laboratories. The building is centrally air­ conditioned and tobacco smoking is prohibited inside. The filters in the air handling unit (AHU) are more effective in filtering large particles (above 10 ~lm) and less effective in filtering smaller particles. The floors are not carpeted, but are covered with ceramic tiles. The floors are generally very clean and maintained regularly. We used a personal air sampler, based on the gravimetric principle, (Models 224-PCXR7 and 224-XR7, SKC, Inc., Eighty Four, USA) to measure indoor TSP concentration (at 2 L min-I). To measure RSP concentration we attached an aluminum cyclone to the sampler (Model 225-01-01, SKC Inc., Eighty Four, USA). The cyclone has a 50% removal efficiency for particle diameters of 5!-lm (dso) at a flow rate of 1.9 L·min- I . Flow rate was measured using the soap bubble technique (±1O%). Teflon filters (37 mm diameter) with pore size l!-lm (model 225-17-01, SKC Inc., Eighty Four, USA) were used after desiccating them with silica gel for 24 h. Filters were weighed in a balance with an accuracy of 10 j.lg. One in every 20 filters was kept as a control blank. Each weighing of the filter was repeated at least twice till a difference of 100 !-lg or less was achieved. Blank corrections were made batch wise. The sampler was placed on a vacant desk on the third floor of the building at a height of 1.5 m above the floor.

Saksena and Vma

Samples were collected twice a week, over 8 h, with the timing roughly matched to the regular work hours of the occupants (09.00-17.00 h). Indoor air monitoring at IHC began in July 1998. Initially only TSP samples were collected indoors, but from April 1999 only indoor RSP samples were collected. We used the method recommended by the Central Pollution Control Board (CPCB) / Indian Standards Institute (lSI) to monitor outdoor particulate matter. A high volume sampler (flow rate: 1m 3.min -1, model HVS, Envirotech, New Delhi, India) were used to sample outdoor levels of TSP using 20 x 25 cm glass fiber filters (Whatman GF/A, Brentford, UK). A similar sampler, with a similar filter, but attached to a cyclone (model APM 460, Envirotech, New Delhi, India) was used to sample outdoor levels of RSP. This sampler configuration has a nominal particle cut-off diameter of 10 /lm. Filters were desiccated for 24 h prior to and after sampling. These were weighed in an electronic balance with an accuracy of 10 j..lg. The flow rates, as indicated by the in-built flow meter, were recorded every hour. The sampler flow meter was calibrated once a month. Outdoor monitoring at IHC began in January 1998. Initially only TSP was sampled, but since June 1999 only RSP was monitored. The samplers were placed at road level, about 50 m north of the Darbari Seth Block, yet within the IHC perimeter. The sampler was 37 m away from the median of Lodi Road. The sampling inlet was at a height of 3 m above the ground level. Samples were collected twice a week over 24h using three 8-hour back-to-back sub­ samples. The batch timings were: 14.00-22.00, 22.00-06.00, and 06.00-14.00 h. One of the aims of the study was to examine the correlation between indoor levels of particulate matter at IHC with outdoor levels at a central fixed site. Delhi has a network of 10 fixed site government monitoring stations. The monitoring station nearest to IHC is Siri Fort, about 4km south. However, at this site CPCB did not collect samples daily during the study period and only provided data for monthly averages in the public domain. The monitoring station at the Income Tax office (ITO), 8 km north of IHC, did collect daily data and provided public access to the data. So we chose to include this site in the present study. The ITO station is considered Delhi's flagship air quality monitoring station, both in terms of the relatively more sophisticated monitoring done at this site and because it is considered a "hot spot". It is a commercial area and the sampler is just by the road. The monitoring methods used by CPCB at ITO for TSP and RSP are the same as described above for IHC for outdoor monitoring.

Indoor Particulate Matter in New Delhi

Monitoring at this site began in 1980s and archived daily data are publicly available June 1999 onwards. We downloaded the daily data of TSP and RSP at ITO from the CPCB website (http://www.cpcb.nic.in/bulletin/ bul.htm). Statistical data analysis was done using SPSS version 12. Outliers were defined as: Q3

+ 1.5 x

IQR

Where: Ql is the first quartile, Q3 is the third quartile and IQR is the inter-quartile range (QrQl)' This definition of outliers is valid for normally distrib­ uted variables. We assumed that all particulate matter concentration variables are distributed log-normally. Therefore, we first log-transformed the data and then applied the above definition to detect, examine and reject outliers. Outliers were detected only in two of the variables: TSP at ITO and indoor RSP.

Results and Discussion

Summary statistics for the concentrations of TSP and RSP at all the indoor and outdoor locations are presented in Table 1. It was observed that outdoor levels of TSP and RSP were higher than the indoor levels. It is noteworthy that while the outdoor particulate matter levels had a coefficient of variation (CV) in the range 42-57%, the CV for the indoor levels was significantly higher, ranging between 80-88%. The variability in the data are also shown in the box plots (Figures 1 and 2). Note that the Table 1. Summary statistics for the indoor and outdoor concentrations of TSP and RSP

n

Mean (~lg.m -3) Coefficient of variation (%) Geometric Mean (!!g.m- 3) Geometric standard deviation

608 468 42

610 186 57

165 315 46

218 195 52

146 232 80

101 119 88

431

160

282

172

170

86

1.48

1.66

1.53

1.65

2.33

2.77

Note: The sample size, n, refers to the number of days.

Indoor Built Environ 2008;17:543-551

545

1000

9

900

8 7

800

'1

700

~ 600 c

'1x

-,.--­ -,-­

6

5

III Q)

"

4

iii

500

> 3

l!! 'E 400

(I)

2

0

:;::l

u

c

0 u 300

o --- ---- -------- ~------- C_ o 100 200 300

24-h stc ndard

200

-'-­

100

0 N=

608 Outdoor TSP at ITO

~ 165 Outdoor TSP at IHC

-,-­

'1

146 Indoor TSP

-..--­

350

-,-­

~ 300 c

o 250

~

'E 200 ~

§ 150 u 100

24-~d

50

--L

o N=

610

218

101

Outdoor RSP at ITO

Outdoor RSP at IHC

Indoor RSP

Fig. 2. Distribution of RSP concentration.

central line in all box plots indicates the median value. The mean outdoor concentrations of TSP and RSP clearly exceeded the 24 h ambient air quality standard prescribed by CPCB (200 and 100 f.!g.m- 3 for TSP and RSP, respectively). Though strictly not comparable, we found that the mean indoor concentrations also exceeded the CPCB standards. Outdoor TSP levels were found to be higher at ITO as compared to those at IRC (which has much less traffic). But the reverse was true for RSP. These differences were statistically significant based on f-tests. It is a common perception that ITO is a traffic "hot spot" [19]. It is also known that in metropolitan areas, finer

546

500

600

Fig. 3. Histogram of indoor RSP with best fit lognormal distri­ bution superimposed.

500

400

400

Concentration (llg·m-3)

Fig. 1. Distribution of TSP concentration.

450

~~_f:_::;: .. _::1_b~~= ........

I _::_"}_ __

Indoor Built Environ 2008;17:543-551

particulate matter such as RSP are related mainly to emissions from traffic. Therefore, the higher levels of RSP at IHC are counter-intuitive. We are unable to speculate about the causes of this observation. Using the Kolmogorov-Smirnov (K-S) test we found that the concentrations of TSP and RSP at all locations had a good fit with a log-normal distribution. For example, Figure 3 shows the distribution of indoor RSP concentrations. In another new and air-conditioned office building in Delhi the indoor levels of TSP and RSP were found to be higher (228 f.!g·m- 3 for TSP and 134 f.!g.m- 3 for RSP) than those observed in the current study [20]. This is very likely due to the much higher outdoor levels of particulate matter observed in that study, as the building was located in a very congested commercial area - Nehru Place. The CV of indoor particulate matter concentration was found to be much less than that we observed. In Hong Kong, the indoor levels of PM IO in mechanically ventilated and air conditioned offices were found to have a mean value of "'50 f.!g.m- 3 [21]. Seasonal variations in particulate matter levels are shown in Figures 4 and 5. The months were classified as follows: November - February (winter), March - June (summer), July - August (monsoon) and September ­ October (post-monsoon). Outdoor levels of TSP at ITO and IRC were found to be highest in winter. This has also been observed by others in Delhi [19,22] and has commonly been attributed to increases in source activity (such as increased traffic) and meteorological conditions such as winter inversions. The lowest levels of outdoor TSP and RSP were observed during the monsoon. This can be attributed to lower source activity levels and wash-out of pollutants from the atmosphere. In contrast to the seasonal patterns observed for outdoor levels, we found that indoor

Saksena and Uma

1100

Table 2. Statistical summary of the RSP/TSP ratio at ITO

1000 900

f"-

800

n

~c

700

Minimum Maximum Mean Coefficient of variation ('Yo) Geometric mean

E

.2 1ii

E

'"

0

c

0 (.)

600 500 400 300

IITSPatiTO

200

IITSPat IHC

100

1.0

0

0.9 0.8

r

g

450

1ii

0.6

~

0.5

400

[l.

~

350

T

0.4

(f)

a:

§ 250

0.3

~

200

0.2

c

150

8

-,­

0.7

500

~ 300

2l

605 0.08 0.80 0.39 33 0.37

0.1

-'­

100 0.0

50

o,,--_......._ _....._ _

~

_ _.......__

Winter

Summer

Monsoon Post-monsoon

Season

Winter Season

Fig. 6. Seasonal variation in RSP/TSP ratio at ITO.

Fig. 5. Seasonal variation in RSP levels.

concentrations for TSP and RSP were highest in summer. Indoor TSP was lowest in the winter and indoor RSP was lowest in the post-monsoon season. We speculate that these indoor seasonal patterns may be caused by changes in (a) number of occupants and their activities and (b) ventilation rates of the air-conditioning system. We were able to calculate the RSPjTSP ratio at ITO because at this site the two particulate matter size fractions were monitored simultaneously daily during the period of this study. We found the mean RSPjTSP ratio to be 0.39 (Table 2). This corresponds to the lower end of the range (0.33--0.66) reported based on India's National Ambient Air Quality Monitoring Programme [23]. Considering the high traffic at ITO and the likelihood of more emissions of smaller size particulate matter associated with vehicles, our study indicates that sources other than traffic continue to be significant contributors to particulate matter levels, even in traffic "hot spots". The RSPjTSP ratio was the highest in winter (Figure 6). This is possibly due to an increase in vehicular activity in the city. The ratio was the

least in summer because of a decrease in traffic and because sources such as dust storms contribute to particles in the higher size ranges. A one-way ANOVA test indicated that seasonal differences in the mean RSPjTSP ratio were statistically significant (F= 40, p< 0.0001). The role of traffic in contributing to RSP is more evident in Figure 7, which shows the difference in the ratios between weekdays and weekends. The weekday ratio of RSPjTSP was found to be statistically significantly higher on weekdays, when there is far more traffic (based on one­ way ANOVA, F= 5, p < 0.02). This was also observed in Kolkata, India in another study [24]. The annual changes in particulate levels at IRC are shown in Figure 8. Outdoor TSP concentrations remained the same between 1998 and 1999 (TSP monitoring was discontinued after that). Outdoor concentration of RSP decreased slightly after 1999 and remained constant after that. We observed a significant reduction in indoor RSP concentrations in 2000 as compared to 1999. Also, we observed that the correlation between indoor and outdoor particulate matter was very weak (Table 3).

Indoor Particulate Matter in New Delhi

Indoor Built Environ 2008;17:543-551

547

1.0 0.9 0.8 0

- -

0.7

t:

1ii 0.6

~

0.5

~

0.4

a.

(f)

II:

0.3 0.2 0.1

- -

0.0 Weekend

Weekday Type of day

Fig. 7. Variation in RSP/TSP ratio at ITO across type of day. 400 II Outdoor TSP ~ Outdoor RSP o IndoorTSP • Indoor RSP

350

300

ef" E

61 250

~

c .2

~

200

Q)

(.)

c 0

(.)

c

'" ::2

150

Q)

100

50

0 1998

2000

1999

2001

Year

Fig. 8. Annual changes in particulate levels at IHe.

This was true irrespective of which outdoor location IHC or ITO - was chosen for analysis. However, the correlation coefficient, r, was just slightly higher for RSP than for TSP, as is to be expected because TSP is very

548

Indoor Built Environ 2008;17:543-551

sensitive to micro-scale local sources. For both TSP and RSP the indoor levels had a better correlation with outdoor levels measured at IHC rather than those measured at the more distant location at ITO. A negative correlation was observed between indoor TSP and outdoor TSP at IHC. We are unable to speculate about the cause of this. At best, variations in outdoor concentration of RSP are able to account for only 25% of the variation in indoor RSP concentration. Given that the overall correlation coefficients were weak, we examined our results to see if in certain seasons the correlations were stronger (Table 4). Though the correlations continued to remain weak, in certain seasons the correlations were higher than those observed based on the pooled data (Table 3). The comparatively higher correlations between indoor and outdoor levels were generally observed during the summer months. As mentioned in the methods section, the indoor concentrations were measured only for 8 h during the day to coincide with normal business hours. On the other hand the outdoor monitoring at IHC and ITO was carried out over 24 h across three consecutive 8-h batches. We, therefore, analyzed the correlations batch-wise (Table 5). Batch-wise outdoor data at ITO was not available, hence the analysis was restricted to the IHC data. Overall, there was no significant improve­ ment in the correlation coefficients compared to the pooled data (Table 3). Contrary to our study, a wintertime study of TSP, PM 10, PM 2 .5 and PM I in 11 offices in Beijing, China found very high correlations of indoor and outdoor concentra­ tions (? ranging from 0.816 to 0.958) [25]. This may be attributed to the fact that (a) the sampling duration was only 2 h in the mornings, (b) these offices were not air-conditioned and (c) windows were kept open during the day. The range of indoor TSP concentration observed in their study was 25-225/-lg.m- 3 and that for PM IO was 1O-170l-lg.m-3. The authors speculated that construction activity and vehicular emissions were the major determi­ nants of indoor particulate matter levels. It is important, also, to remember that their design was cross-sectional rather than longitudinal. The study of another air-conditioned building in Delhi [26] found the longitudinal indoor--outdoor correla­ tion coefficient, r, ranging from -0.19 to 0.44 for TSP and -0.6 to 0.44 for RSP across floors. In a study of shops and nurses' dormitories in Bangkok correlation coefficients between indoor and outdoor PM lO were found to range between 0.41 to 0.69 [27]. Srivsatava and Jain [28] studied the indoor-outdoor relationship

Saksena and Vma

Table 3. Correlation between indoor and outdoor particulate matter

TSP at ITO

Pearson Correlation N

RSP at ITO

Pearson Correlation

TSP at IRC

Pearson Correlation

RSP at IRC

Pearson Correlation

TSP indoor

Pearson Correlation

RSP indoor

Pearson Correlation

N N N N N

I 608 0.736** 605 0.866** 32 0.683** 209 0.027 87 0.194 84

0.736** 605 I 610 0.868** 32 0.704** 211 0.135 88 0.460** 84

0.866** 32 0.868** 32 I 165 (a) 0 -0.490** 40 0.037 14

0.683** 209 0.704** 21l (a) 0 1 218 -0.032 38 0.503** 37

0.027 87 0.135 88 -0.490** 40 -0.032 38 I 146 (a) 0

0.194 84 0.460** 84 0.037 14 0.503** 37 (a) 0 1 101

**Correlation is significant at the 0.01 level (2-tailed).

aCannot be computed because at least one of the variables is constant.

Table 4. Season-wise correlations among the outdoor and indoor particulate levels

Winter

Summer

Monsoon

Post-monsoon

TSP at ITO RSP at ITO TSP at IRC RSP at lEC TSP indoors RSP indoors TSP at ITO RSP at ITO TSP at IRC RSP at IRC TSP indoors RSP indoors TSP at ITO RSP at ITO TSP at IRC RSP at IRC TSP indoors RSP indoors TSP at ITO RSP at ITO TSP at IRC RSP at IRC TSP indoors RSP indoors

1 0.713** 0.594* 0.590** 0.389* 0.220 1 0.692** (a) 0.669** -0.122 0.131 1 0.556** 0.862 0.440* -0.509 -0.399 1 0.861 ** 0.800** 0.759** -0.376 -0.420

0.713** 1 0.568* 0.468** 0.685** 0.469** 0.692** I (a) 0.766** 0.217 0.527** 0.556** 1 0.245 0.651 ** -0.582* 0.266 0.861 ** 1 0.839** 0.767** -0.400 -0.364

0.594* 0.568* I (a) -0.193 -0.276 (a) (a) 1 (a) -0.881 -0.057 0.862 0.245 1 (a) 0.387 -1.000** 0.800** 0.839** 1 (a) -0.721** -1.000**

0.590** 0.468** (a) 1 -0.121 0.584 0.669** 0.766** (a) 1 -0.049 0.609* 0.440* 0.651 ** (a) I 0.975** 0.519 0.759** 0.767** (a) 1 -0.890 -0.841

0.389* 0.685** -0.193 -0.121 1 (a) -0.122 0.217 -0.881 -0.049 1 (a) -0.509 -0.582* 0.387 0.975** I

(a) -0.376 -0.400 -0.721** -0.890 1 (a)

0.220 0.469** -0.276 0.584 (a) 1 0.131 0.527** -0.057 0.609* (a) 1 -0.399 0.266 -1.000** 0.519 (a) 1 -0.420 -0.364 -1.000** -0.841 (a) 1

**Correlation is significant at the 0.01 level (2-tailed). *Correlation is significant at the 0.05 level (2-tailed). aCannot be computed because at least one of the variables is constant.

of TSP longitudinally in a university building (not air-conditioned, 24-h samples) and cross-sectionally at other residences and commercial areas (3-h samples) in Delhi. They observed indoor TSP levels of 217Ilg.m-3 in the university building and 383 J..Lg.m- 3 in buildings in

commercial areas. They observed values of ,2 between indoor and outdoor TSP levels of 0.30 at the university building and 0.85 in the commercial areas. In a university building in Chongju, Korea mean levels of indoor PM 2 .5 were found to be 25Ilg·m-3 and a strong

Indoor Particulate Matter in New Delhi

Indoor Built Environ 2008;17:543-551

549

Table 5. Indoor-outdoor correlations on a 8-hour basis

Indoors TSP

Pearson Correlation Sig. (2-tailed) N

Indoor RSP

Pearson Correlation Sig. (2-tailed) N

Outdoor TSP at IHC -0.446** -0.521** -0.337* 0.004 0.001 0.034 40 40 40 Outdoor RSP at IHC 0.390* 0.422** 0.503** 0.009 37

0.002 37

0.025 33

**Correlation is significant at the 0.01 level (2-tailed). *Correlation is significant at the 0.05 level (2-tailed).

correlation with outdoor levels of PM 2.5 was found (r 2 = 0.785) based on eight days of monitoring with 24-h sampling periods [29]. Baek, Kim and Perry [30] studied the indoor-outdoor relationship cross-sectionally in offices in Seoul and Taegu, Korea. The mean values of RSP indoors and outdoors were found to be 99 Jig.m- 3 and 78 Jig.m- 3, respectively. In contrast to our study they found indoor RSP levels to be higher in winter. They too found weak correlations between indoor and outdoor levels of RSP in offices (r = 0.17 and 0.25 in summer and winter, respectively). In homes the correlations were higher, with "r" ranging from 0.63 to 0.76. The correlations among the outdoor concentrations at IHC and ITO were strong, r=0.866 for TSP and 0.704 for RSP (Table 3). These correlations were higher in the summer and post-monsoon months (Table 4). This is one of the longest longitudinal studies of indoor-outdoor relationships of particulate matter carried out in a developing country. The mean values of indoor TSP and RSP concentrations, though much higher than those observed in similar micro-environments in developed countries, were less than the outdoor concentrations, but higher than the Indian ambient standards. However, while in developed countries studies (and in some developing countries studies) strong longitudinal correlations have been observed between indoor and outdoor concentrations, results of this study were contrary. The variation in outdoor concentra­ tion of RSP could explain only 25% of the variation in indoor concentration. The day-to-day variation in indoor concentration in terms of the coefficient of variation was

550

Indoor Built Environ 2008;17:543-551

more than 80%-· much higher than that observed outdoors. This strongly supports the belief that indoor­ outdoor relationships cannot be generalized but are strongly dependent on factors such as type of building, ventilation, climate, etc. An implication of this study is that the daily outdoor particulate matter concentration based on central fixed sites cannot serve as a reliable surrogate for the daily personal exposure of office work­ ers. Consequently, the health effects models developed based on the time-series approach are less applicable to adult office workers and perhaps more applicable to population sub-groups such as infants and the elderly. This has also been observed in Europe in the EXPOLIS study [31] and in the US [32]. The present study was conducted in a new and sophisticated centrally air­ conditioned building. It is likely that the filtration process reduced the impact of outdoor pollution. On the other hand El-Shobokshy and Hussein [33] observed in offices in Riyadh, Saudi-Arabia, that sometimes air-conditioning can increase indoor particulate levels if the filters get overloaded and are not maintained well. Thornburg et al. [34] too have observed that air-conditioning systems can increase indoor particulate levels in commercial buildings because of the air exchange rate, in the absence of efficient filters. In other office buildings with natural ventilation, homes, schools, etc., the higher air exchange rates and lack of mechanical filtration may lead to stronger indoor-­ outdoor correlations. To counteract this, the presence of unique indoor sources in these types of less affluent buildings may lead to weaker indoor-outdoor correla­ tions. Considering that in developing countries there is not enough evidence to suggest that indoor source activities remain comparatively constant, as also the ventilation, etc., on a day-to-day basis, there is a need to do further research in a wide variety of buildings, homes and schools to understand the real determinants of the indoor-outdoor air pollutant relationships, especially for particulate matter. Harmonization of monitoring protocols is very desirable, especially in terms of monitoring duration, frequency and the degree of synchronization between indoor and outdoor monitoring. Acknowledgment

The authors wish to thank Mr Ved Prakash of TERI, Delhi, for his assistance in data collection. We also thank the authorities at TERI for their financial support of the ambient air quality monitoring program at IHe.

Saksena and Uma

..

References I Pope CA, Dockery DW, Schwartz J: Review of epidemiological evidence of health-effects of particulate air-pollution: Inhal Toxicol 1995;7(1):1-18. 2 Anderson HR. Ponce de Leon A, Bland JM, Bower JS, Strachan DP: Air pollution and daily mortality in London: 1987-92: Br Med J (Clinical Research Edn.) 1996;312(7032):665-669. 3 Schwartz J: Air pollution and daily mortality in Birmingham, Alabama: Am J Epidemiol 1993;137(10 (print»:1136-1147. 4 Schwartz J: Air pollution and hospital admis­ sions for the elderly in Birmingham, Alabama: Am J Epidemiol 1994; 139(6 (print»:589-598. 5 Verhoeff AP, Hoek G, Schwartz J, van Wijnen JH: Air pollution and daily mortality in Amster­ dam: Epidemiol 1996;7(3 (Print»:225-230. 6 Pope CA, 3rd, Dockery DW, Spengler JD, Raizenne ME: Respiratory health and PMIO pollution. A daily time series analysis: Am Rev Respiratory Dis 1991;144(3 Pt I (Print»:668-674. International Scientific Oversight 7 HEI Committee; Health effects of outdoor air pollution in developing countries of Asia: a literature review. Special Report 15. 2004: Health Effects Institute: Boston. 8 Spengler JD, Soczek ML: Evidence for improved ambient air-quality and the need for personal exposure research: Environ Sci Technol 1984;18(9):A268-A280. 9 Sexton K, Spengler JD, Treitman RD: Personal exposure to respirable particles - a case-study in Waterbury, Vermont: Atmos Environ 1984;18(7):1385-1398. 10 Spengler JD, Treitman RD, Tosteson TD, Mage DT, Soczek ML: Personal exposures to respirable particulates and implications for air-pollution epidemiology: Environ Sci Technol 1985;19(8):700-707. II Wallace L: Indoor particles: A review: J Air Waste Manage Assoc 1996;46(2):98-126. 12 Vtell MJ, Samet JM: Particulate air pollution and health. New evidence on an old problem: Am Rev Respiratory Dis 1993;147(6 Pt 1 (print»: 1334-1335. 13 Janssen NA, Hoek G, Brunekreef B, Harssema H, Mensink I, Zuidhof A: Personal sampling of particles in adults: relation among personal, indoor, and outdoor air concentrations: Am J Epidemiol 1998;147(6):537-547.

Indoor Particulate Matter in New Delhi

14 Lawrence AJ, Masih A, Taneja A: Indoor/ outdoor relationships of carbon monoxide and oxides of nitrogen in domestic homes with roadside, urban and rural locations in a central Indian region: Indoor Air 2005;15(2):76-82. 15 Kumar P: Characterization of indoor respir­ able dust in a locality of Delhi, India: Indoor Built Environ 2001;10(2):95-102. 16 Khillare PS, Pandey R, Balachandran S: Characterisation of indoor PM sub(lO) in residential areas of Delhi: Indoor Built Environ 2004;13(2):139-147. 17 Saksena S, Singh PB, Prasad RK, Prasad R, Malhotra P, Joshi V, Patil RS: Exposure of infants to outdoor and indoor air pollution in low-income urban areas - a case study of Delhi: J Exposure Anal Environ Epidemiol 2003; 13(3):219-230. 18 Sharma DN, Sawant AA, Vma R, Cocker I, David R.: Preliminary chemical characteriza­ tion of particle-phase organic compounds in New Delhi, India: Atmos Environ 2003;37(30):4317-4323. 19 Aneja VP, Agarwal A, Roelle PA, Phillips SB, Tong Q, Watkins N, and Yablonsky R: Measurements and analysis of criteria pollu­ tants in New Delhi, India: Environ Int 2001;27(1):35-42. 20 Saksena S, Prasad RK, and Shankar VR: Daily exposure to air pollutants in indoor, outdoor and in-vehicle micro-environments: a pilot study in Delhi: Indoor Built Environ 2007;16(1):39-46. 21 Lee S-C, Guo H, Li W-M, and Chan L-Y: Inter-comparison of air pollutant concentra­ tions in different indoor environments in Hong Kong: Atmos Environ 2002;36(12):1929-1940. 22 Monkkonen P, Vma R, Srinivasan D, Koponen IK, Lehtinen KEJ, Hameri K, Suresh R, Sharma VP, and Kulmala M: Relationship and variations of aerosol number and PM 10 mass concentrations in a highly polluted urban environment­ New Delhi, India: Atmos Environ 2004; 38(3):425-433. 23 TERI; Review of past an on-going work on urban air quality in India. 2001: Tata Energy Research Institute: New Delhi. p. 101. 24 Karar K, Gupta AK, Kumar A, Biswas AK, and Devotta S: Statistical interpretation of weekday/weekend differences of ambient

particulate matter, vehicular traffic and meteorological parameters in an urban region of Kolkata, India: Indoor Built Environ 2006; 15(3):235-245. 25 Liu Y, Chen R, Shen X, Mao X: Wintertime indoor air levels ofPMIO, PM2.5 and PMI at public places and their contributions to TSP: Environ Int 2004;30(2):189-197. 26 Prasad RK, Vma R, Kansal A, Gupta S, Kumar P, and Saksena S: Indoor air quality in an air-conditioned building in new Delhi and its relationship to ambient air quality: Indoor Built Environ 2002;11(6):334-339. 27 Tsai FC, Smith KR, Vichit-Vadakan N, Ostro BD, Chestnut LG, and Kungskulniti N: Indoor/outdoor PMIO and PM2.5 in Bangkok, Thailand: J Exposure Anal Environ EpidemioI2000;IO(l):15-26. 28 Srivastava A, Jain VK: Relationships between indoor and outdoor air quality in Delhi: Indoor Built Environ 2003;12:159-165. 29 Lee HS, Kang B-W, Cheong J-P, Lee S-K: Relationships between indoor and outdoor air quality during the summer season in Korea: Atmos Environ 1997;31(11):1689-1693. 30 Baek S-O, Kim Y-S, Perry R: Indoor air quality in homes, offices and restaurants in Korean urban areas-indoor/outdoor relation­ ships: Atmos Environ 1997;31(4):529-544. 31 Kousa A, Oglesby L, Koistinen K, Kunzli N, Jantunen M: Exposure chain of urban air PM2.5-associations between ambient fixed site, residential outdoor, indoor, workplace and personal exposures in four European cities in the EXPOLIS-study: Atmos Environ 2002;36(18):3031-3039. 32 Sarnat JA, Koutrakis P, Suh HH: Assessing the Relationship between personal particulate and gaseous exposures of senior citizens living in Baltimore, MD: J Air Waste Manage Assoc 2000;50(7):1184-1198. 33 El-Shobokshy MS, Hussein FM: Correlation between indoor-outdoor inhalable particulate concentrations and meteorological variables: Atmos Environ 1988;22(12):2667-2675. 34 Thornburg J, Ensor DS, Rodes CE, Lawless PA, Sparks LE, Mosley RB: Penetration of particles into buildings and associated physical factors. Part I: Model development and com­ puter simulations: Aerosol Sci Technol 2001;34(3):284-296.

Indoor Built Environ 2008;17:543-551

551