Subscriber access provided by - Access paid by the | UCSB Libraries
Article
Characterizing the Spatial and Temporal Patterns of Open Burning of Municipal Solid Waste (MSW) in Indian Cities Ajay S Nagpure, Anu (Anuradha) Ramaswami, and Armistead G. Russell Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.5b03243 • Publication Date (Web): 08 Oct 2015 Downloaded from http://pubs.acs.org on October 13, 2015
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 31
Environmental Science & Technology
Characterizing the Spatial and Temporal Patterns of Open Burning of Municipal Solid Waste (MSW) in Indian Cities Ajay Singh Nagpure1*, Anu Ramaswami1, Armistead Russell2 1
Center for Science, Technology, and Environmental Policy, Hubert H. Humphrey School of Public
Affairs, University of Minnesota, Twin Cities, MN, USA 2
School of Civil and Environmental Engineering, Georgia Institute of Technology, Atlanta, GA, USA
∗
Corresponding author phone: +1-919-800-8194; E-mails:
[email protected];
[email protected] Word count: ~5195 words (excluding references), 5 figures and 1 table
1 ACS Paragon Plus Environment
Environmental Science & Technology
1
ABSTRACT
2
Open-burning of municipal solid waste (MSW) is a major source of PM emissions in developing
3
world cities, but few studies have characterized this phenomenon at the city and intra-city
4
(neighborhood) scale relevant to human health impacts. This paper develops a consistent field
5
method for measuring the spatial frequency of the incidence of MSW-burning and presents
6
results in three neighborhoods of varying socioeconomic status (SES) in Delhi, India, observed
7
in winter and summer over 2 years. Daily MSW-burning incidents ranged from 24-130/km2-day
8
during winter and 5-87/km2-day during summer, with the highest intensity in low SES
9
neighborhoods. Distinct seasonal and diurnal patterns are observed. The daily mass of MSW-
10
burned was also estimated at 90-1170 kg/km2-day and 13-1100 kg/km2-day in highest to low
11
SES neighborhoods, in winter and summer, respectively. The scaled-up estimate of total MSW-
12
burned for Delhi city ranged from 190-246 tons/day, about 2%-3% of total generated MSW;
13
morning-burning contributed >65% of the total. MSW composition varied systematically across
14
neighborhoods and season. Agra had much higher MSW-burning (39-202 incidents/km2-day;
15
672-3485kg/km2-day) in the summer. The field method thus captures differences in MSW-
16
burning across cities, neighborhoods, diurnally and seasonally, important for more fine grained
17
air pollution modeling, and for tracking/monitoring policy effectiveness on-ground.
18
Keywords:
19
Emissions
Municipal solid waste burning; Urban waste management; Air pollution; PM
20 21 22 23
2 ACS Paragon Plus Environment
Page 2 of 31
Page 3 of 31
Environmental Science & Technology
24
INTRODUCTION
25
Environmental scientists and pollution control agencies are increasingly focusing on air pollution
26
by particulate matter (e.g., PM10, particles whose diameters are less than 10 µm, and PM2.5),
27
because short- and long-term exposure to these pollutants is associated with a number of health
28
impacts, including respiratory and cardiovascular disease, adverse birth outcomes, and cancer1-10.
29
Based on these studies and evidence from occupationally exposed populations, the International
30
Agency for Research on Cancer (IARC) has identified PM2.5, specifically, as a Group 1 human
31
carcinogen11. Most recently, estimates from the Global Burden of Disease (GBD) study indicate
32
that outdoor particulate pollution is the fifth major cause of premature death and disability-
33
adjusted life years lost in India after high blood pressure, indoor air pollution, tobacco smoking
34
and poor nutrition, with about 695,000 premature deaths estimated per year12. Furthermore, when
35
the outdoor air pollution sources are concentrated near residential areas, the related emissions
36
can infiltrate indoors, contributing to the indoor air-quality problems. In addition to these serious
37
health effects, PM pollution has also been associated with aesthetic damage of culturally
38
important statues and monuments in Indian cities, notably the Taj Mahal in Agra13.
39 40
Open burning of municipal solid waste (MSW) is increasingly being recognized as a major
41
source of PM10 and PM2.5 emissions in the cities of developing countries14-16, along with other
42
sources, such as traffic emissions and industrial and power plant combustion operations. MSW
43
refers to non-industrial and non-medical solid waste generated in municipal (urban) areas, i.e.,
44
garbage generated from households and commercial establishments in cities. Open burning of
45
MSW (MSW-burning) refers to MSW burned in urban neighborhoods, without the use of
46
incinerators, with the resulting combustion-related pollutants released to the atmosphere at the 3 ACS Paragon Plus Environment
Environmental Science & Technology
47
surface, with relatively less buoyant plumes. MSW-burning differs from biomass burning, as the
48
latter traditionally has been used in the context of burning crop residues and other biogenic
49
residues (such as dung-cakes), often in agricultural settings. In contrast, MSW is not purely
50
biomass, but often includes plastics, rubber and metal-containing refuse, the burning of which
51
releases toxic emissions (including halogenated compounds from plastics burning). All these
52
factors highlight the importance of MSW-burning to public health at the urban and intra-city
53
scale.
54 55
In emission inventories conducted for a few major Indian cities, the Central Pollution Control
56
Board of India17 estimates that open MSW-burning may contribute 5% to 11% of all direct PM
57
emitted from sources within a city boundary. Similar estimates are reported for Mexico City
58
(3% to 30%) and Ulaanbaatar, Mongolia (4% to 7%)15,18. Thus the phenomenon of MSW-
59
burning is likely to be very important to urban public health in India. A majority of Indian cities
60
with populations exceeding 1 million rank among the top 100 world urban areas with worst PM10
61
pollution19. A recent global study16 highlights the importance of MSW-burning, estimating that
62
it could contribute approximately 8% and 22% of direct PM emissions in India and China,
63
respectively. Despite the emerging recognition of the importance of this phenomenon, most city-
64
scale air pollution emission inventories neglect MSW-burning because of the large uncertainty in
65
the estimates of the fraction of MSW being burned and the paucity of actual field observations of
66
the phenomenon at the city scale. Without any field estimation or baseline study, the
67
implementation and success of any laws or policies attempting to remedy the situation are
68
difficult to evaluate, as are the potential health impacts on different populations within the cities.
69
4 ACS Paragon Plus Environment
Page 4 of 31
Page 5 of 31
Environmental Science & Technology
70
At present, mass estimates of MSW-burning vary widely and are typically derived from either
71
expert opinions or rule-of-thumb assumptions that estimate the burn fraction of daily MSW
72
generated (Table S1). These estimates of MSW-burning vary and typically are not based on
73
detailed analysis of the specific locations or the relationships to neighborhood characteristics
74
within a city. Furthermore, these estimates do not account for the spatial variation in MSW-
75
burning emissions and their linkage to socioeconomic status (SES) within the cities. Given that
76
MSW-burning emissions have their main impact on human health at the local scale
77
(neighborhood and city), developing a robust bottom-up field approach to studying MSW-
78
burning within cities (at intra-urban scales) and its variation according to neighborhood SES and
79
infrastructure provisions is important for assessing health and local environmental impacts.
80 81
To date, there have not been detailed field observations exploring the spatial and temporal
82
patterns of MSW-burning at the urban and intra-urban scale.
83
established for characterizing MSW-burning patterns, with sensitivity to infrastructural and
84
socioeconomic differences among cities and neighborhoods. Such methods are important for
85
beginning to understand the linkage between neighborhood characteristics and MSW-burning
86
emissions, and are essential for assessing health impacts and designing location-sensitive
87
programs and policies to reduce the practice of MSW-burning. This is important because
88
exposure to emissions from MSW-burning, and their related health effects, occur near-field,
89
meaning that urban residents in proximity to MSW-burning are likely to be more at risk. The
90
incidence of MSW-burning is likely to show sharp gradients across SES, with higher incidences
91
observed in lower income areas with less infrastructure provision. Yet little is known about these
5 ACS Paragon Plus Environment
Field methods are not yet
Environmental Science & Technology
92
finer scale intra-urban variations that have important implications for health and urban inequity
93
in exposure to pollution.
94 95
The objective of this paper is to develop a consistent field method for measuring the daily
96
incidence (i.e., the spatial frequency) of MSW-burnings and to estimate the mass of MSW
97
burned per unit area in different neighborhoods of varying socioeconomic status (SES) in Indian
98
cities.
99 100
We first present general observations about the MSW-burning in Delhi, India during early field
101
work conducted in 2014. We then present a method for field measurement of the spatial
102
frequency of MSW-burning incidents, MSW composition, and mass of MSW burned at the intra-
103
urban scale. The method was tested in Delhi in winter and summer 2014-15 in three different
104
neighborhoods of varying socioeconomic status, with and without MSW collection services. The
105
method was then streamlined and deployed in Agra in summer 2015, offering an inter-city
106
comparison.
107 108
Key questions asked during the development of this method are: What is the best way to track
109
and represent the daily spatial frequency of MSW-burning in a neighborhood, i.e., whether by
110
the number of incidents per unit of area, or per household? Does the method capture seasonal/
111
temporal patterns, and is it repeatable? Can the field method developed to record MSW-burning
112
incidence also be applied to assess the composition and mass of the MSW being burned? What
113
are the ways that neighborhood scale MSW-burning data can be scaled-up to the whole city to
114
contribute to a city wide emission inventory?
6 ACS Paragon Plus Environment
Page 6 of 31
Page 7 of 31
Environmental Science & Technology
115
Developing methods for estimating the mass of MSW generated and the fraction that is burned at
116
the neighborhood scale, and its relationship with various factors such as the provision of MSW
117
collection services (or lack thereof), as well as the socioeconomic status of households (HHs),
118
can help ground-truth the rules of thumb currently being used. This can help inform better
119
policies to address the issue of MSW-burning at the city and neighborhood scale.
120 121
METHODS
122
Transect distance sampling approach
123
A distance sampling approach, also called the line transect method20, was employed to estimate
124
the spatial frequency, volume, and composition of MSW-burning at the neighborhood and city
125
scale. The distance sampling approach draws upon a field method typically used by researchers
126
for estimating the biological density and/or wildlife abundance in a specific habitat. In this
127
method the observer moves along a line/road and counts the objects in a predetermined distance
128
from the transect line (typically just visible range is used as the distance). Object density is then
129
estimated by the total object count and surveyed area (Fig. S1). Fundamentally, it is assumed that
130
all objects within the transect line of sight are detected. Employing a similar approach, a transect
131
line (route) was designed for each neighborhood that covered a large proportion of each
132
neighborhood (about 12-40%, See Fig. 1), and included both residential and commercial areas
133
with diverse street types. Each MSW-burning incident along the transect route is recorded by the
134
field researcher. The transect route, as well as the latitude and longitude waypoint of each MSW-
135
burning incident were recorded by a hand-held Garmin GPS 72. Each neighborhood transect in
136
Delhi was repeated three times each season, during summer and winter, in 2014 and 2015. The
137
transects were covered by vehicle (two-wheeler or auto rickshaw) and by walking in some
7 ACS Paragon Plus Environment
Environmental Science & Technology
138
neighborhoods where the streets were not conducive to motorized travel. Several cross-Delhi
139
transects spanning multiple neighborhoods were also conducted (Fig. S3). For an intercity
140
comparison, multiple-neighborhood, cross-city transects were conducted in the nearby city of
141
Agra during summer 2015.
142 143
Delhi (28.61° N, 77.20° E, population 16.7 million) is the capital city of India and covers an
144
area of 1,484 km2. Delhi is divided into 2,064 neighborhoods which are called colonies in India
145
21,26
146
houses the world heritage historical monument the Taj Mahal.
. The city of Agra (27°12' N 78°12' E) has an area of 141 km2 with 1.63 million people and
147 148
Daily patterns and spatial frequency of MSW-burning
149
Pre-transect observations of MSW-burning incidents were conducted in Delhi along three
150
neighborhood transects, on two hour intervals throughout the day and late evening. These
151
observations indicated that there were more MSW-burning incidents in the morning compared to
152
the evening, typically starting around 6 am. MSW-burning was not observed in the afternoon,
153
and then commenced again at dusk. Therefore the formal transect studies tracked both morning
154
and evening observations each day, but not in the afternoon. Transect sampling stopped at 10 pm
155
due to safety concerns for the researcher. However, any ash piles from burning past 10pm were
156
noted by the researcher during the next morning’s transect and were added to the morning tally.
157
Daily temperatures were also recorded on each of the transect study days. Individual transects
158
were repeated over three days each in each of the three neighborhoods and in the three cross-
159
Delhi transects, both in January 2014 and in May 2014 (pre-monsoon). The same transects were
160
repeated in January and May 2015. Thus three different neighborhood transects and three cross-
8 ACS Paragon Plus Environment
Page 8 of 31
Page 9 of 31
Environmental Science & Technology
161
Delhi transects were observed over the period of 2 years. The team learned from the 2014 field
162
experience, and recorded both morning and evening transects consistently every sampling day in
163
2015.
164 165
Neighborhood choice & context
166
The transect studies of MSW-burning were conducted in three neighborhoods of Delhi, that
167
varied significantly in socioeconomic status and levels of MSW collection service.
168 169
Delhi’s Municipal Valuation Committee Report21 has classified the neighborhoods of Delhi
170
according to different parameters such as the capital value of land, rental value, age of the
171
colony, condition of the road, quality of physical and social infrastructure, and economic status.
172
Neighborhoods are categorized into 8 types (A-H) with A representing the highest SES and H
173
the lowest. As per the report, Brijpuri (BP) belongs to the F and G category, Jangpura Extension
174
(JPE) to the B category, and Safdarjung Enclave (SJE) to the A category. The BP neighborhood
175
was selected to represent a low-income slum area with poor MSW collection. The two high
176
income areas, JPE and SJE, with apparently similar MSW service levels, were selected to assess
177
any differences in MSW-burning patterns across more wealthy neighborhoods. The
178
characteristics of the three neighborhoods, and their MSW-management infrastructures, are
179
summarized below (and in Table 1 and Fig. S4).
180 181
Brijpuri (BP) is considered a slum, with a low income level, a very high population density of
182
about 87,000 people/km2, and minimal MSW collection services. Also, due to the low income of
183
the residents, there were no privately employed household waste collectors. Only one
9 ACS Paragon Plus Environment
Environmental Science & Technology
Page 10 of 31
184
municipality-employed street sweeper was seen by the research team; street sweepers are tasked
185
with only sweeping debris from the streets. No designated waste collection centers - which are
186
used for discarding MSW and serve as a collection point for truck to haul away the waste - were
187
observed during the canvassing of Brijpuri in January 201422. Residents also informed the
188
research team that a city-provided waste-pickup truck comes once a day, though it cannot access
189
a majority of the narrow streets22.
190 191
Safdarjung Enclave (SJE) is considered a high-income area, based on census data26 and also
192
supported by visual observations of the housing types in the area. It is categorized as a type “A”
193
neighborhood, reflecting the best SES in Delhi21. Households in SJE employ household waste
194
collectors who collect and transport the MSW by tricycle to neighborhood waste collection
195
centers. A typical waste collection center
196
collection centers are noted in Supplementary Fig. S4b), and the municipality trucks haul away
197
waste once or twice a day. In addition, several municipal street sweepers serve the area and were
198
observed by the research team.
serves 500-600 homes (locations
of the waste
199 200
Jangpura extension (JPE) is considered a high- to medium-income neighborhood, with a slightly
201
lower neighborhood rating of “B”21 compared to SJE. SJE and JPE have lower population
202
densities than BP, approximately 14,000-20,000 people/km2. JPE households also employ
203
household waste collectors and have waste collection centers located in the neighborhood (1 for
204
every 500 to 600 HHs) (Location of the waste collection centers are noted in Supplementary Fig.
205
S4c).
206
10 ACS Paragon Plus Environment
Page 11 of 31
Environmental Science & Technology
207
Composite MSW generation data were gathered in the three neighborhoods (Table 1). The data
208
were collected by enlisting the services of the neighborhood household waste collector (except in
209
BP where the researcher did the collection as no waste collector is employed by the households).
210
MSW generated by 25 households was collected by the waste collector according to their normal
211
routine. All MSW generated by these homes was aggregated and then segregated according to
212
the type (biodegradable & recyclable) and weighed by the researcher, in a way similar to the
213
method given by Ojeda et al., 200837. Dividing the composite trash weight by the number of
214
homes served by the waste collector yielded the average waste generated per household. This
215
process was repeated on 3 separate days.
216 217
The average mass of the total MSW generated per person per day in BP, SJE, and JPE was 0.17
218
kg, 0.39 kg, and 0.39 kg, respectively. The percentage of the MSW that was
219
biodegradable/compostable for BP, SJE, and JPE was 90%, 75%, and 78%, respectively.
220 221
In the SJE and JPE neighborhoods, at least 13% to 45% of the neighborhood area was covered
222
by the road transect, as shown in Fig. 1, to ensure representative coverage. The BP transects
223
were done along two routes; one accessible by two-wheeler vehicle, and another route was only
224
wide enough for pedestrians (the researchers walked this transect). Each of the transects covered
225
approximately 3 to 6 km, and the same transects were used for estimating seasonal variability in
226
winter and in summer.
227
information was recorded: the GPS location, as well as the estimated weight, composition, and
228
any apparent reason of burning (e.g. if people were warming themselves it was assumed to be
229
due to the need for heat).
During the sighting of each MSW-burning incident, the following
11 ACS Paragon Plus Environment
Environmental Science & Technology
230
For cross-Delhi transects, similar records of incidents of MSW-burning were recorded in
231
different parts of the city, including the central core. The cross-Delhi transects covered diverse
232
neighborhoods ranging from A-G, and included commercial areas and ring roads. About 40km of
233
distance was covered in each cross-Delhi city transect (Fig. S3). Data have been also collected
234
from Agra during summer 2015, by laying down two transects that ranged from 35 to 45 km and
235
covered diverse neighborhoods of the city.
Page 12 of 31
236 237
Daily MSW-burning incidences per unit area and per unit household
238
The number of MSW-burning incidents per unit of road length was converted to per unit area
239
using the average line of sight distance on both sides of the road (Fig. S1), which was equivalent
240
to the distance between the centerline of the transect road and its parallel road. The data were
241
also converted to number of incidents per household by dividing the area by the housing density
242
in each neighborhood (after subtracting for green spaces).
243 244
Composition & coarse estimation of the mass of MSW burned
245
Since it is difficult to weigh burning MSW, samples of MSW were obtained and weighed before
246
and during different stages of burning, by sprinkling water on it and subsequently subtracting the
247
added water weight. This coarsely estimates the mass of the burning MSW, while its
248
composition was assessed by spreading out the burning MSW (after extinguishing) and recording
249
its partially burned components. In this manner, an average weight and composition per burning
250
incident could be estimated from the observation of each MSW-burning incident, during this first
251
deployment of the field method.
252
12 ACS Paragon Plus Environment
Page 13 of 31
Environmental Science & Technology
253
Scale-up from transect studies
254
Data were scaled up to the city level for Delhi using the following different approaches:
255
Approach (i). MSW-burning incidents per household scaled by neighborhood type:
256
In this method we have classified all the neighborhoods of Delhi into two Groups (1) E-H:
257
Lower SES, and, (2) A-D: Higher SES, according to the categories given by the valuation
258
committee21. The scaling-up was done on the basis of the number of households in the different
259
neighborhood types. Incidence of MSW-burning per household from BP were applied to
260
estimate the MSW-burning from Group (1) neighborhoods, while the same data from SJE were
261
used to calculate estimates for Group (2) neighborhoods. Our follow-up study of the social
262
actors engaged in waste management (details are reported in 22), found that
263
management is better in JPE than in the other two neighborhoods because of the direct
264
involvement of the local residents and their neighborhood association in overseeing the daily
265
waste pickup operations. Since such active management of MSW collection is uncommon, we
266
considered JPE to be atypical, and therefore opted to focus on SJE and BP as representative of
267
high- and low-SES areas, respectively. Our scale-up estimate using this approach is a first
268
approximation of a land-use informed scale-up of the observed neighborhood MSW-burning.
269
Future studies may consider traversing multiple neighborhoods of different types, for improved
270
aggregation by neighborhood classification. Neighborhood routes include the main roads and
271
arterials leading to them, as well as associated commercial establishments present locally with
272
the households.
MSW
273 274
Approach (ii) Cross-city transect MSW-burning incidents and mass per area: In the second
275
approach, multiple neighborhoods are traversed in well-planned cross-city transects. Such
13 ACS Paragon Plus Environment
Environmental Science & Technology
276
transects should be planned carefully to cover diverse areas of the whole city. The average of
277
multiple cross-city transects reported per unit area is then scaled to the total urban area23 of the
278
city 920 km2 for Delhi and 185km2 in Agra.
Page 14 of 31
279 280
In general, we consider the first approach, i.e., of knowing and sampling neighborhoods of
281
varying SES and scaling up by the population residing in the neighborhood types, to be a better
282
approach. This first approach is expected to converge with the second approach, as long as the
283
cross-city transects give a representative sample of diverse neighborhoods of the city.
284 285
RESULTS
286
MSW-burning spatial frequency: differences across neighborhoods (Intra urban)
287
MSW-burning frequencies per unit area were found to be highest in the low-SES neighborhood
288
(BP), and in the morning in all neighborhoods during both winter and summer (Fig. 2a). As seen
289
in Fig. 2a, the spatial frequency of total daily MSW-burning incidents (morning plus evening)
290
ranged from 24 incidents/km2-day to 130 incidents/km2-day during winter and 5 incidents/km2-
291
day to 87 incidents/km2-day during summer, and were found to be highest in the low SES
292
neighborhood (BP). Field results showed that differences in daily MSW-burning incidents per
293
unit area between the neighborhoods were considerable, with JPE (24 incidents/km2-day) having
294
about 87% less incidence frequency than BP (130 incidents/km2-day), based on full-day
295
observation in winter (Fig. 2a). Similar patterns were observed in summer with the difference
296
between JPE (5 incidents/km2-day) and BP (87 incidents/km2-day) being about 94%. Likewise,
297
the estimated daily mass burned per unit of area (Fig. 2b) was also highest in BP, where
298
infrastructure services are poor.
14 ACS Paragon Plus Environment
Page 15 of 31
Environmental Science & Technology
299 300
The results from JPE and SJE (higher-SES areas) show, as expected, lower MSW-burning
301
incidence (Fig. 2a) and lower mass burned per unit area (Fig. 2b) and per households (not
302
shown) than BP (statistically significant at a = 0.05). It is important to note that the spatial
303
frequency of MSW-burning incidences was non-zero in these high SES areas that have waste
304
collection services. More surprising is that JPE had statistically lower MSW-burning incidence
305
during winter compared to SJE, which has a higher SES and a higher neighborhood
306
classification. Both have similar levels of MSW service provision. A follow-up survey of
307
households (described in ref 22) indicated that factors such as the participation of neighborhood
308
associations in monitoring waste collection service provides are important in reducing MSW-
309
burning incidents22. A summary of these findings reported in Table 1 indicates, for example, that
310
65% of the households surveyed in JPE were aware of the neighborhood association and its role
311
in informally monitoring waste management in the neighborhoods, compared to less than 5% of
312
homes in SJE. Furthermore, the neighborhood association in SJE did not take on a waste
313
management role. Details of the social actor study are provided elsewhere22. These observations
314
(both physical and social) indicate that the field method was able to detect these finer differences
315
among neighborhoods quite consistently in different years.
316 317
Daily and seasonal patterns
318
Daily and seasonal variation in the spatial frequency of MSW-burning incidences was also
319
significant. For the diurnal patterns, the MSW burning frequency observed in the evening was
320
always lower than the morning in all three neighborhoods, in summer and winter (See Fig. 2).
15 ACS Paragon Plus Environment
Environmental Science & Technology
321
No MSW-burning incidents were observed in JPE and SJE during the evening transects in
322
summer.
Page 16 of 31
323 324
Seasonal variations in MSW-burning were also observed with winter always higher than
325
summer. SJE experiences the largest differences in daily MSW-burning rates between the winter
326
(56 incidents/km2-day) and the summer (6 incidents/km2-day), while BP has the smallest
327
difference between winter (130 incidents/km2-day) and summer (87 incidents/km2-day).
328 329
In terms of the seasonal patterns in the estimate of daily MSW-burned per unit area, the low-SES
330
BP neighborhood has the largest daily mass of MSW burned (1170 kg/km2-day), recorded in
331
winter. This is about 14 times higher than JPE (89 kg/km2-day) full day MSW-burning in winter.
332
For summer the full-day mass MSW-burning in BP is estimated at about 1100 kg/km2-day,
333
followed by 18 kg/km2-day in SJE and 13 kg/km2-day in JPE (Fig. 2b). There is no statistical
334
difference between the daily mass burnt per unit area observed in BP during summer (1170±108
335
kg/km2-day) and winter (1100±481 kg/km2-day). This suggest that lack of collection services in
336
low SES BP results in similar daily MSW mass burned per unit area, both in summer and
337
winter.
338 339
Another interesting note that we also did not observe a significant difference (a = 0.05) in total
340
daily MSW-burning incidents between JPE and SJE in summer. Our field observations suggest
341
that SJE showed higher night-time MSW-burning than JPE in winter, when household guards
342
were often observed burning MSW to keep warm. Restrictions by residents in JPE may have
343
limited this practice, resulting in winter-time differences between these two neighborhoods.
16 ACS Paragon Plus Environment
Page 17 of 31
Environmental Science & Technology
344
When MSW-burning occurs to provide warmth, often firewood and other combustibles are
345
added to the fire, as noted below in the composition details discussed next.
346 347
Composition of MSW Being Burned
348
Field observations indicate (Fig. 3) that the composition of the MSW being burned varies
349
significantly and systematically across the three neighborhoods and across season. In winter,
350
added purchased firewood is seen burning in the wealthier neighborhood to provide heat, while
351
paper and plastics were found burning in lower-SES areas for the purpose of providing heat. The
352
burning of plastics is particularly hazardous24 and reflects an environmental justice issue in the
353
slum areas. These observations show that significant amounts of the materials like plastic and
354
purchased firewood, which are not present in household waste streams, are being burned with
355
MSW in winter. In summer, compostable waste (both kitchen waste and leaves/twigs) are
356
dominant in waste burning in all the neighborhoods. Yard waste is burned more dominantly in
357
wealthier areas, while kitchen waste is predominant in lower-SES areas. The summer inter-city
358
comparison showed remarkable similarity in the waste composition between the low-SES area of
359
Delhi and Agra in summer. The waste composition data suggest that much of the waste can be
360
composted in summer. Neighborhood composting could be considered as an alternative to the
361
present system.
362 363
Scale-up from Transect Studies
364
Applying the two approaches, we estimate the average daily spatial frequency of MSW-burning
365
incidents in Delhi to be 38-40 incident/km2-day in winter and 25-31 incident/km2-day in summer
366
(Fig. 4a) and the average daily mass of MSW-burned per unit area in Delhi to be 236-263
17 ACS Paragon Plus Environment
Environmental Science & Technology
Page 18 of 31
367
kg/km2-day in the winter and 206-246 kg/km2-day in the summer (Fig. 4b). The Morning
368
burning accounting for more than 65% in both seasons (Fig. 4). It is interesting to note that the
369
averages scale up similarly in Delhi, using both approaches, which is expected when the cross-
370
city transects sample a number of diverse neighborhoods. The daily total mass of MSW burned
371
for the whole city is estimated to be about 190 to 246 tons/day. These preliminary calculations
372
estimate that ~2-3% of MSW generated (8390 tons/day) 25 is burned in city streets in Delhi. The
373
estimated percentage burned of total generated waste are within the range specified in the various
374
literatures (Table S1) for Indian megacities, but it appears to be not as large as the higher
375
percentages of 10% assumed by others. However, we note that burning in landfills may also
376
contribute further.
377 378
In contrast, the inter-city comparison showed Agra to be very different, with significantly higher
379
MSW-burning spatial frequencies and a higher mass/area compared to Delhi (Fig. 4). For Agra
380
the MSW burned was observed to be much higher in three different SES areas that were covered
381
– ranging from frequencies of 39 to 202 incidents/km2-day, and 672 to 3485 kg/km2-day in
382
highest- to lowest-SES neighborhoods in summer 2015 (Fig.5). This compares with