DISCUSSION

Water quality analysis
The higher values of ionic concentrations at JK, ML, RC, KM, SM, TH and KN and in contrast, low values at VN, RM and UL explained variation in water chemistry and sewage as the sole source of increased cations and anions inflow to wetlands. The EC was found to be conservative surrogate with strong correlation with aquatic biota such as diatom communities (Soininen, 2004). The accumulation of more ions has also lead to increased demand for oxygen (COD and BOD) which was evident at JK, ML, TH and KT. Poor management in discharging of untreated domestic, industrial and commercial waste in to the water bodies and land use pattern in catchment area were the observed major environmental stressors that affected the water quality (Beyene and Legesse, 2005). Furthermore in PCA analysis, PC3 revealed high BOD and COD concentrations at sewage fed, polluted wetlands such as ML, TH inlets and UL. VN, SM and KN were separated from rest of the sites indicating no influence of any of the variables as explained in PCA plot.

Diatom and macroinvertebrate distribution
Species distribution, richness and evenness reflected significant changes across sites and gradients in water chemistry. Diatoms comprised species indicating pollution levels such as N. palea, D. confervaceae and C. meneghiniana which are known to be tolerant species with high saprobic number (van Dam et al. 1994) and hence reflected pollution at JK and ML. Corixidae, pollution indicator group among macroinvertebrates was also evident of increasing degradation in water quality at JK and ML. The biota is found to have been influenced by both the primary (physical and chemical water variables) and secondary (human disturbances and land use in catchment) factors (Nautiyal and Verma, 2009). Species richness was less in case of macroinvertebrates when compared to diatoms (Figure 4) because of the response of former to poor habitat quality. It has been reported that less habitat-dependency of diatoms when compared with macroinvertebrates consents to more occupancy of diatoms on several habitats (McCormick and Cairns 1994; and Reid et al. 1995).

Diatom and macroinvertebrate indices:
Most of the water quality variables (BOD, COD, EC, ALK and chlorides) revealed significant positive correlation with %PT and TDI indices indicating pollution level. The increasing TDI with increasing BOD, COD and EC, the representative of ionic concentrations, gave evidence for the relationship between physical and chemical variables with the biological variables (Atazadeh et al.2007). Implementing GDI indices could not be useful for urban wetlands though the genus level identification explains pollution, it may cover up species contribution, its tolerance and sensitivity in indices calculation. The resolution of this index is considerably lower than the TDI indices which rely on species level identification. However, this index may be useful for the purposes of providing an initial indication that an aquatic system is polluted but not its severity.

The low correlation coefficients between macroinvertebrate metrics and environmental parameters is caused by the fact that benthic communities are not always affected by this kind of pollution and considerable diversity can be present in places with heavy chemical loads (De Pauw and Heylen, 2001; Beyene and Legesse, 2005). On the other hand several report scores for macroinvertebrate metrics correlating strongly with a wide variety of water quality parameters, including toxicants. The families such as Belostomatidae, Diptera, Gerridae, Hydrometridae, Notonectidae, Gerridae, Pleidae, Hydrophiloidea, Thiaridae and damsel fly lacks tolerance scores in both FBI and BI (Mandaville, 2002) which are documented as pollution sensitive groups (Crane et al. 2000). Though Thiaridae was dominant, VK outlet site had 0 indices score because the tolerance value of Thiaridae is not specified in either of the indices. Similar problem was recorded with families such as Belostomatidae and Gerridae at Jakkur, Pleidae at Mallathally and Diptera at Rachenahalli. However, the BI index includes most of the families than the FBI index and can be further modified to use as indicators.

The interconnected wetlands explain the significance of environmental conditions for species occupancy. High ionic level is attributable for 56% of CMEN taxa at Jakkur, while it was 35%, at Rachenahalli with less ionic level compared to former. The species composition reflected not just the water chemistry but physical habitat quality such as substrate composition, turbulence and land use patterns as factors that influence diatom macroinvertebrate assemblages (Giller and Malmqvist, 1998).

The clustering of sampling sites based on the relative abundance of diatoms and macroinvertebrate revealed the trophic status and scale of organic pollution across Bangalore. The sites categorized under class I and II were relatively clean waters with fewer nutrients and less disturbed wetlands. The sites which are grouped in the class III to IV are inflow sites with sewage and industrial effluents while, class I and II include most of the outlet sites. The oligotrophic status of outlet sites explains the role of ecosystem in uptake of nutrients by macrophytes and other organisms. According to TDI, most of the sites come under mesotrophic classification, while inlets of ML and JK were clustered as Meso-eutrophic to eutrophic (Class IV-V) with high constituents of EC, CHL, hardness and organic nutrients, which were also evident from water quality and dominance indices.

In macroinvertebrates, the family viz., Nepidae and Belostomatidae, being dominant at JKO1 and Thiaridae at JKI1, are underprovided with tolerance values and though with elevated ionic values, grouped as good water quality as per BI index. The tolerance values for these macroinvertebrates are not defined in BI and FBI, hence indices fail to distinguish the appropriate trophic status and subsequently a slight difference between diatom and macroinvertebrate indices classification was recorded. JK and ML wetlands, where high EC, BOD and COD influencing water quality was considerable due to intense human disturbances, these sites trophic status was not reflected through BI indices. The autecological values of individuals explaining trophic status has to be standardized to support the development of macroinvertebrates as bioindicators in Indian condition. The application of a number of biotic indices and scores leads often to conflicted results (Iliopoulou-Georgudaki et al. 2003). Several studies as well, Triest et al. 2001, Gonçalves et al. 2008, pointed out in their comparative study that primary producers - diatoms are more indicative for water quality and trophic status than macroinvertebrates. Lack of species ecological data could be the reason for less prevalence of macroinvertebrates as indicators.

It has been reported earlier that macroinvertebrates as poor water quality bioindicators while diatoms as better in response to pollution load. Nevertheless, the species composition and environmental variables vary among temperate and tropical regions, and leaves an unknown margin of uncertainty as differences in environmental conditions and species sensitivity are not taken into account in the latter case (Rico et al. 2011). The diatom indices developed based on taxa of temperate regions was successful in reflecting pollution in tropical ecosystems as diatoms are cosmopolitan (with few exceptions) compared to macroinvertebrates due to lack of species data and are specific to tropical region (Mandaville, 2002). Diatoms and macroinvertebrates provided consistent information on water quality assessment and were suggested for use of combination of indices as biological indicators of the water quality of temperate regions (European Parliament and European Counseling, 2000; Soininen and Könönen, 2004; Schletterer et al. 2011). On the other hand the studies on tropical ecosystems explained macroinvertebrate composition as liable to severe physical habitat and chemical water quality degradation. Thus, macroinvertebrates are less diverse and are not found in tropical, severely polluted sites with low dissolved oxygen levels and diatoms are more reliable and indicate a gradient of pollution than macroinvertebrates (Samways, 2005; Beyene et al. 2009). This suggests the implementation of indices and integrating region specific taxa and methods for improved biomonitoring.