Methods

Diatom and water sampling
Diatom samples were collected from a total of 106 sampling sites of 43 wetlands of Bangalore watershed region during February-April 2009 and February-April 2010 following guidelines as per standard protocols (Karthick et al.2010). Post monsoon period was chosen to avoid the diluted water quality and the studied wetlands and sampling codes are listed in Annexure 1. The wetlands and sampling sites were prioritized based on the earlier studies carried out by Bengaluru Bruhat Mahanagara Palike (BBMP). Thirty wetlands are situated in the densely built-up areas (residential and industrial areas) characterized by waste inflow and less water inflow. The remaining thirteen wetlands were situated in less populated regions at the city’s periphery.

Diatom sampling was carried out covering all possible habitats viz., plants (epiphytic diatoms), stones (epilithic diatoms) and sediments (epipelc diatoms). Samples were cleaned following hot HNO3 and HCl method. Slides of cleaned samples were prepared using Naphrax (refractive index, r.i. 1.74) as mounting agent and 400 frustules were counted under the Olympus light microscope BX–51 with DIC at 1000×. Taxa were identified up to the species level (and not up to variety level) in order to avoid ‘force fit’ identification of tropical taxa to European and North American taxa. Identification was performed by referring the standard key literatures (Krammer & Lange-Bertalot 1991, Lange-Bertalot 2001, Taylor et al. 2007b). Water samples were collected along with diatom samples for analyzing environmental variables from various points. Water temperature (WT), dissolved oxygen (DO), total dissolved solids (TDS), salinity (SAL), pH and electric conductivity (EC) were measured on site using EC500 pH conductivity meter (Extech, Massachusetts). Nutrient concentrations, namely, nitrates (N) and phosphates (P), major ions  chlorides (CL), total hardness (TH), calcium hardness (CaH), magnesium hardness (MgH), alkalinity (ALK), chemical oxygen demand (COD) and biological oxygen demand (BOD) were analyzed in the laboratory by following APHA (2005).

Data analysis

Diatom taxa occurring in at least one sample with a relative abundance >10% were considered for further statistical analyses to understand inter and intra distributional pattern and its relationship with water conditions. Detrended canonical correspondence analysis (DCCA), a direct multivariate gradient analysis, was performed using CANOCO for windows 4.5 (ter Braak 1986) to evaluate the relationship between environmental variables and diatom species composition patterns (or distribution). DCCA attempts to remove arch effect in samples and spreads the points to avoid a clumping pattern. Two-way indicator species analysis (TWINSPAN) was performed through TWINSPAN for Windows (WinTWINS) version 2.3 developed by Hill & Šmilauer (2005).  This software considers the classification of samples first and then the classification of species based on its ecological preferences. TWINSPAN classifies the sampling points according to the relative abundance of species and defines diatom community types based on the reciprocal averaging (Gauch 1982). It performs well in the classification of benthic assemblages when compared with other clustering techniques.