Methods

The city of Bangalore is situated between 12°39’00’’ to 13°13’00’’ N and 77°22’00’’ to 77°52’00’’ E covering an area of 741 km2.  The entire study area was divided into 250grids of size 2.7 km × 2.7 km.  Unauthorized dumping of USW occurs both inside as well as outside of the city boundary.  Field investigations of dump sites were studied separately in three segments, namely, a) inside core area of the city, b) the region between the core area and the city boundary (outer city) and in c) peripheral area of the city upto a buffer distance of 10 km from city boundary (Figure 1).  This distance of 10 km was chosen based on the earlier studies carried out in Bangalore [16, 17].  The surveys were carried out during pre and post monsoon periods of August 2010 to April 2012.

Identification of dump site locations:

The first part of survey was focused on maximizing the identification of unauthorized dump site locations.  This method involved the participation of 45 trained student volunteers riding on motor-cycles.  These volunteers were provided Google Earth (GE) images marked with probable dump sites.  Volunteers were provided an initial orientation and hands-on training before undertaking the survey.  They were trained in the usage of GPS, identification of wastes and estimation of their composition and determining the size of dump sites visually by taking them to a known dump site and trying out the technique learnt.  This ensured that there was a good match of recorded data among volunteers and compared the observations they made and allowed corrections to be made to become able to assess the length and width of dumps visually with a reasonable level of accuracy while also ensuring personal safety.  Visual survey techniques are commonly used for estimation of target organism abundances in ecological field studies conducted in terrestrial [28, 29, 30] and aquatic environments [31, 32].  This approach facilitated covering a large area in a short time.  It could be carried out with a basic level of training and was less expensive.  This ensured that nearly all the unauthorized dump sites were surveyed and their locations identified and verified with reasonable accuracy.  The approach and method included two stages. 

In first stage, the purpose of site observation was to find the locations of unauthorized waste dump sites by traversing on motorable roads and paths in the designated grid.  Each of the student volunteer team (2 persons) was allocated 2 to 4 grids to be completed in 2 days of survey time.  All the dump site discovered were photographed and its location determined using a calibrated GPS.  Observations were recorded on hard copy of map (GE images of the respective grids).  The USW spread area was determined by visually estimating the length and breadth of the dump while waste composition was determined based on visual observations.  This approach could identify the predominant waste type in dump sites.  Rapid visual estimate was adopted considering the local constraints including hostile attitude of the section of society dependent on waste for livelihood.  The track points of volunteers were later verified using GE and compiled data were verified through field visit by an 'expert' to maintain accuracy of the assessment.  

In second stage, identified locations and observations were verified in many ways.  At the end of the survey of each grid, all GPS and physical data was downloaded and location or grid accuracy were validated by superimposing it on GE and estimates of composition were verified based on the photographs.  These waste dumps were classified into six different categories based on the predominant form of wastes found during survey, namely – plastics, organic, construction debris, indeterminate (fresh wastes mixed up making segregation difficult), old (older dumps with traces of decomposed domestic USW and absence of most of the organic wastes) and other (non-USW wastes often from small industry, upholstery, etc.) based on waste type.  After the survey of all the grids were completed (about 3-5 months) each of these participating groups submitted a report and made a presentation of their observations.  Random field verification of the locations for both accuracy of location, size and general composition were carried out by experts.  From a total of 696 dump locations identified by volunteers, an expert group randomly verified more than 40% (269) of identified locations distributed across all the four quadrants of the city.  The above total of 269 locations comprised of 193 sites lying within the core city and 76 sites lying in the region outside the core city boundary.  On the outskirts of the city at 33 selected locations the area of the dump site was physically measured using calibrated step measurement (SM) as used in step transect method [33] and these locations were also measured for their area using GE.  Physical measurement of dump area was done through the measurement of dump length and width wherever these dumps were predominantly rectangle in shape or physically and socially accessible.  Locations suffering continuous dumping over 2 to 3 years could be identified in GE (spatial resolution 0.5-30 m).  Hence, comparison of expert based area estimate versus volunteer estimates was therefore done predominantly with long term older dump sites to ensure greater accuracy. 
Monitoring of locations
A second survey was carried out after the completion of the first year survey by an expert team.  A total of 452 locations visited by student volunteers during the first year was revisited.  This was conducted to monitor whether the dumps identified in the previous year were continued.  New dump locations in the peripheral area outside the city were also identified. 

Data Analysis:

Regression analysis was carried out for all the USW dumps whose spread area were determined by SM with the dump spread area estimated by other methods such as, GE or Visual estimation (VE) to assess the probable relationships [19].  Mean and standard deviation were also estimated for different waste categories.  In addition to this, a paired “t” test was conducted to determine grid-wise change in number of dump locations that had occurred in two consecutive years of field survey.  All statistical analyses were performed with PAST software [34].