2.1. Study area and its characteristics

The study was undertaken in two major wetlands in the Bangalore city namely Bellandur and Varthur wetlands (Fig. 1). Bangalore has the distinction of having interconnected wetlands due to undulating terrains with three major watersheds (Ramachandra et al., 2016). Bellandur Lake is the largest Lake in the Bangalore city and spreads across an area of 365 ha (mean depth 1.8 m). It is situated 5 km upstream of Varthur Lake. These lakes receive ~500 million litres per day (MLD) of untreated or partially treated wastewater, which include municipal wastewater from i) residential areas near the old Bangalore Airport, and ii) regions around Challaghata and Koramangla that directly flows to these lakes through connected drains. Varthur wetlands/lake is also situated in the south of Bangalore, covers a water-spread area of 220 ha (mean depth 1.1 m) built for catering the decentralised demand of water for domestic and agricultural uses. The Varthur-Bellandur Lake catchment has seen large-scale land use changes after 2000, consequent to the rapid unplanned urbanization process in the region. The characteristics of both the water bodies are provided in Table 1.

Table 1 Characteristics of the study area.

2.2. Water sampling and analysis

Water samples from inlets, middle and outlets were collected across different seasons to evaluate the influent and the effluent water quality. These lakes had a varying extent of floating macrophytes during different seasons, which impeded the use of boats for sampling. Only samples closer to the shore could be reliably sampled at specific times of a day as the wind induced drift of floating macrophytes on the lake made time-specific sampling of all the points unfeasible.

The average annual precipitation of Bangalore is about 700 - 850 mm and temperatures vary from 14 °C (December to January) to 33° C (maximum during March to May). There are two rainy periods, i.e. south-west monsoon (June to September) and north-east monsoon (November to December) (Mahapatra et al., 2011a; Ramachandra et al., 2016). During these periods, fresh water enters the lake as runoff. Water samples were collected regularly every month from five predetermined sampling points (locations were recorded using a hand-held pre-calibrated GPS (Global Positioning System, Garmin 48 and 60) to represent inlets, outlets and midpoints (Fig. 1). Physico-chemical parameters - p^H, air temperature, water temperature, TDS, EC, turbidity, transparency and DO were measured at site following the standard protocol. Water samples were collected in 1 litre disinfected containers for estimation of chemical parameters in the laboratory (APHA AWWA WEF, 1998). Table S1 in the Supplementary material lists parameters and method adopted for analysis.

2.2.1. Water and key biota analysis

Water samples were also collected seasonally for 15 months from inflows, middle reaches and outflows of Bellandur and Varthur Lakes (Fig. 1) to examine the influent and the effluent water quality together with capturing the water quality in the middle reaches. Nutrient removal was calculated as per equation (1), for COD, BOD, N-species and P-species to assess the treatment efficiencies (in %).

2.2.2. Bacterial, algal and macrophyte analysis

100 ml of sample were collected from the select locations and were fixed with 70% alcohol. Species were identified using light microscope (Lawrence and Mayo) at 40 * with the help of morphological keys as per literatures (Prescott, 1954, 1962; Desikachary, 1959). The algal members were enumerated by following earlier protocols (Mahapatra et al., 2013a,b). The bacterial count and morphology assessments were done through flow cytometry (FACS Caliber) following Gasol and Giorgio (2000) and with Electron Microscopy. These results were compared with the results of light microscope observations. Flow cytometry technique was adopted for rapid enumeration apart from exploration of different bacteria population, which is superior to plating as it is prone to contamination. Bacterio-plankton population was analysed by first filtering the samples with 2.5 mm sieve and then were analysed by the FSC (forward scatter) plots. Macrophytes were collected from the sampling locations and were identified following standard keys for freshwater plants in India by CDK Cook, 1996.

2.2.3. Bacterial analysis through scanning electron microscopy (SEM) and flow cytometry (FC)

At a stable p^H, the cells were fixed in 2.5% glutaraldehyde. Samples were dried after dehydration through a series of ethanol in buffer of increasing strength (30, 50, 70, 80, 90, and absolute 10 min each). Specimen were mounted, gold sputtered and examined through electron microscopy following earlier protocols (Mahapatra et al., 2014). 1.5 ml of 2 micron filtered samples (wastewater/sludge supernatant) were fixed with 1% paraformaldehyde + 0.05% glutaraldehyde (final), allowed in the dark, deep frozen in liquid nitrogen for 10 min to fix and then stored frozen at -70 ° C for estimating the bacterial abundance. FACS Caliber of Becton and Dickinson with a laser emitting at 488 nm, samples were treated at low speed (approx.18 ml/min) with acquisition of data in log mode until 10000 events. Samples were diluted when the sample acquisition rate is higher than 800 cells/s.Usually 10ml per 200-ml sample is added of a 10⁶ /ml solution of yellow-green 0.92 mm Polysciences latex beads as an internal standard. Bacteria were detected by their signature in a plot of Side scatter (SSC) vs. Forward scatter (FSC). Adjustments were made in the settings so that the beads fall in channels ~10 ³ for SSC and FL1/ FSC as per standard protocol (Giorgio et al., 1996; Gasol and Giorgio, 2000).

2.3. Data analysis

Relationship between changes in physico-chemical variables and biological variables (relative abundance of algal classes and algal/bacterial abundance) was assessed through the computation of nonparametric Spearman's rank correlation coefficient (r). Multivariate analysis - canonical correspondence analysis (CCA) was carried out to understand interplays between biological communities with varying physicochemical variables and seasons and to know relationships among them (Martin-Cereceda et al., 2001; Iscen et al., 2008). Higher variability in abundance was handled through log₁₀ transformation of the values. The spatial similarity and patterns across sites was assessed through Bray - Curtis cluster analysis (CA). An open source statistical package PAST 2.14 was used to implement statistical analyses.