Impacts of a massive flood event on the physico-chemistry and water quality of river Pampa in Western Ghats of India

ABSTRACT Kerala state (India) experienced a devastating flood event during the month of August 2018 which has brought considerable changes to the environment and ecosystem, especially to rivers. One of the worst affected basin was Pampa river which experienced a severe flood. Aims of this investigation were to define the quality of Pampa river water after the flood. The analysis of physico-chemical parameters indicated transitory variations from pre-flood status as reported by former workers. The pre-flood neutral to the acidic property of water shifted towards alkaline conditions throughout the river stretch post flood. Alkalinity and salinity was almost twice the pre-flood scenario, though the values were within the ranges required for fish survival and growth. Decreased conductivity and increased dissolved oxygen were also the major changes observed post flood. The study indicates that this large-scale flood resulted in a considerable reduction in the concentration of heavy metals and nutrients in the water. Based on water quality index, spatially Pampa river water can be categorised from ‘Good’ to ‘Unsuitable’ category. The cluster analysis distinctly differentiates the middle stream stations from lower stream stations. PCA/FA loadings indicate that electrical conductivity, alkalinity, total hardness, calcium, and PO4 3 – P show strong positive loadings which indicate that salinity factor and natural weathering are the main factors controlling the water quality. This information would be practical to forecast the health of the ecosystem after the flood and also to develop adequate management plans.


Introduction
With the rising need of water for drinking and agro-industrial developments, the requirement of freshwater is increasing exponentially [1]. Among the various natural sources of freshwater, rivers are the most important. However, with intense anthropogenic activities and rising events of climate change, quality of river water is constantly deteriorating [2]. Climate plays a vital role in the ecosystem processes as the hydrological regimes in freshwater ecosystems, especially the riverine ecosystems, are strongly influenced by the climate change [3]. The climatic variations are found to affect the rainfall patterns and the number of rainy days. The floods caused by heavy rains along the course pick up soil, waste materials and other pollutants which finally gets deposited into the river. Hence, the chances of contamination of freshwater bodies are very high during and immediately after the flood which deteriorates the quality of the water [3].
In August 2018, Kerala, the southern state of India experienced one of the worst flood in the century, which caught the attention of researchers and policymakers [4]. The extreme flood in Kerala caused due to the intense rainfall affected millions of people, both socially and economically. Preliminary estimates indicate that the flood caused an economic loss of around 3 USD billion and loss of human life of 440 numbers [5]. In addition, there has been extensive damage to terrestrial and aquatic ecology due to change in course of the rivers, siltation and accumulation of debris [6]. This incident warranted an urgency to assess the impact of the unprecedented serious event on the rivers of Kerala. The Kerala state is blessed with 44 rivers and the holy river Pampa, the third-longest (176 km) in the state which supports significant fishery, was the worst hit during the monsoon flood. Hence, an urgent need was felt to understand the impact of the flood on the water quality of the river.
Study on water chemistry is essential to determine the suitable use of any water body for various purposes like drinking, irrigation, recreation, as well as fisheries and hence continuous monitoring and assessment is necessary. Water quality index (WQI) [7] assessment plays an important role in valid interpretation of the large dataset on various physico-chemical parameters which is combined and transformed into clear and exploitable information. Floods can alter the water quality of rivers to a great extent according to earlier reports [8,9]. Though WQI has been used to evaluate the quality of water in various rivers across India (Ganga [10], Brahmaputra [11], Cauvery [12], and Godavari [13]) and the world over (Soan [14], Nile [15], Murucupi [16], and Yacuambi [17]), minimal information is available on the WQI of Pampa river and also for the impacts of the flood on WQI of river water. Multivariate analysis viz. cluster analysis (CA), principal component analysis (PCA) and factor analysis (FA) is widely used in recent years to interpret the large volume water quality dataset [18][19][20]. All these studies indicate the appropriateness of multivariate statistical techniques in identifying the sources of pollution and also in grouping the monitoring stations possessing similar characteristics. In most of the studies these techniques were used in a large area or large river for assessing the pollution trend and source. Nevertheless, systematic methodical studies in this angle for a small river in the Western Ghats are scarce and this method along with WQI may help to increase the level of objectivity and certainty.
Flooding is one of the most well-known events for contributing to water contamination [21] and water quality of the river is subject to significant change during this extreme event. However, there is a lack of data on the dynamics of water quality during the flood and afterwards due to difficulty in sampling following the aftermath of such a disaster [22,23]. Therefore, it is imperative to understand the water quality after the flood and each such extreme event represents an opportunity for emergency preparedness and preventive measures to improve. Several previous studies [24][25][26][27] reported the water quality of the Pampa river. No, study of water quality during flooding has been conducted previously in the Pampa river, Kerala and this is the first study aimed at determining the water of a river immediately after a flood event. Several previous studies [24][25][26][27] on Pampa river water quality are also lacking in the major information on the suitability of this freshwater for various utilities. Investigations on the water quality characteristics in the post-flood scenario are very vital in understanding the flood phenomena and its socio-economic impacts. Hence, the present investigation was taken up with the aim (1) to analyse the quality of Pampa river water after the August 2018 flood incidence in Kerala and compared with the pre-flood status of Pampa river as reported by earlier researchers [24][25][26][27], (2) to evaluate the efficiency of multivariate statistical techniques in interpreting river water quality and pollutant interactions, and (3) to investigate the spatial differences in the water quality monitoring stations for identifying potential contamination source. The study also evaluated the WQI to determine the use of Pampa river water for variousdesignated purposes after the flood. Pearson's correlation matrix was used to determine the interrelation between various physico-chemical properties of water. PCA has been applied for simple interpretation of large dataset and also CA was employed for grouping the different monitoring stations depending on similarity in quality parameters.

Study area
Pampa river flowing through Kerala, the southern state of India has a length of 176 km and a catchment area of 2235 km 2 [28]. It originates at Pulachimalai hill (Idukki district) in the Western Ghats at an altitude of 1,650 msl, drains through four major districts of Kerala before joining Vembanad Lake. Average drainage density of Pampa basin is 2.96 km/km 2 with an annual average discharge rate of 3.9 km 3 /year [29]. The river falls under the tropical humid climate with two distinct monsoons, the South-West (May/ June to August) and the North-East (October-November) monsoon. The river stretch based on its physiographic nature can be divided into three stretches, upland, midland and lowland.
Middle and lower stretches of the river basin were selected for this study as the entry to upper reaches was restricted due to heavy flood-induced damages. The middle and lower stretches were divided into seven segments for studying the impacts of the flood on the river. The stations are depicted in Figure 1 and a brief description of each station is given in Table 1.

Water sampling and analysis
The post-flood water samples (triplicates) were collected immediately after the flood event at 30 cm depth from the middle of the river at each station and stored in an appropriate condition for further analysis.

Calculation of Water quality index (WQI)
Weighted arithmetic method [7] was used to evaluate the WQI of Pampa river following standard limits as prescribed by BIS [31] and WHO [32]. Eleven important parameters that significantly control the water quality were considered for the calculation of WQI by using following steps: Step 1: Here, q n indicates the quality rating, V n and S n are the observed and standard prescribed limits, respectively, for the n th water quality parameter, whereas V i is the ideal value which is theoretically '0' for all parameters except pH and dissolved oxygen for which the ideal values are 7 and 14.6, respectively.
Step 2: The unit weight (W n ) was calculated by a value inversely proportional to the recommended standard value S n of the corresponding parameter.
Where k is the proportionality constant and was calculated as follows: Step 3: The WQI was calculated by dividing the product of quality rating and unit weight by summation of all the unit weights for the corresponding water quality parameter, following the equation,

Multivariate statistical analysis
PCA was done to identify the source of pollution [33,34]. The eigenvalues generated during PCA analysis measures the significance of the PCs, and factor loadings determine the correlation between the PCs and original variables. The hierarchical Cluster Analysis (CA) was used for grouping the different sampling stations based on the similarity of the water quality parameters. Pearson's correlation (r) analysis was carried out to determine the association between the various parameters. SPSS version 20 software was used to analyse the dataset.

Physico-chemistry of water
Physico-chemical indicators of Pampa river water during the post-flood period are given in Table 2 and compared with the pre-flood status as reported by previous researchers [24][25][26][27] and shown in Table S1. The average depth of the river was >5 m at all the stations, except Veeyapuram ( Table 2). As there is no bibliography on the water depth of the Pampa river, it is impossible to compare pre and post-flood conditions. Higher transparency ranging from 0.36 m to 1.3 m ( Table 2) was observed in the middle stretch of Pampa river as compared to the values (0.135 m-0.195 m) reported by Rajan and Anila [27] for the same stretch (Table S1). High depth of water after a flood may lead to this higher transparency. In lower stretches, the overall transparency was higher as compared to middle stretch. Shoreline macrophytes in lower stretches play a definite role in the higher transparency regime, as they prevent soil erosion as well as trap suspended silt in their leaves [35].
Water temperature is one of the important parameters for the aquatic ecosystem since all the chemical indicators of water quality are influenced by it. Due to the cold-blooded nature, body temperature of fish changes if there is a small variation in temperature which in turn influences their growth and metabolism [36]. The temperature at different stations ranged between 25.5°C and 32°C ( Table 2). As compared to previous reports [25,26] (Table  S1), water temperature during post-flood was low, particularly at the middle stretch. This may be because of the rainwater with a lower temperature that got mixed with the river water [37]. However, Sajudeen et al. [24] reported similar values as in the present study. Spatially, the water temperature varies from one site to another which may be attributed to the geographical location. Being a tourist spot, downstream stretches had shown higher temperature which may be due to the merging of effluents with high temperature.
Fish can tolerate water temperature from 30°C to 35°C [38]; however, the range for the survival of most of the fish species is 24°C -30°C [39]. The variations in temperature of river water are influenced by the rainfall and anthropogenic activities which in turn affects the survival and migration of the fish. The high-intensity flood can negatively affect the river fish biomass and productivity [36]. However, the drop in temperature in this study is a positive sign. The pH is another important parameter and most of the fish can survive in the pH range 6.5-9.5 while the range between 7.5 and 8.5 is ideal. The pH level (7.4 to 8.0) of the present study is in the desirable range (Table 2); however, the water has become moderately alkaline in Pampa river, whereas the previous studies reported [24][25][26][27] acidic to neutral nature (Table S1). In general, heavy rainfall lead to acid precipitation and thus leaching of acid to waterbodies which make them acidic in nature. Contradicting this, the pH in river Pampa has increased during the study period which could be attributed to the mixing of wastewater during the flood. In general, most of the rivers in Kerala originating from the Western Ghats is slightly acidic to moderately alkaline in nature due to the status of the catchment [40,41]. Mohan et al. [9] also found a slightly higher pH in the Periyar river in Kerala after the August 2018 flood. However, none of the recorded pH crosses the prescribed limit (8.5) of WHO, and BIS for drinking water.
DO value also showed a spatial variation. Maximum DO was observed at Ranni (9.68 mg/l) and the lowest at Kainakary station (4.6 mg/l). In general, water is well oxygeniated in all the stretches, though decreased towards the river mouth. A comparison was made with reference to the previous studies (Table S1) which indicate that there is only negligible change in the DO at middle stretches; however, a drastic reduction was observed at the river mouth. This low value after the flood events may be due to the discharge of high pollutant loads (organics and nutrients) along with the surface runoff. Report [42] suggests that minimum DO requirement for supporting various fish species ranges between 4 mg/l and 5 mg/l. The values obtained in the present study showed that the Pampa river water was well oxygenated even after the flood which is sufficient for the survival of the biotic community. The DO value also indicates that the river is less polluted especially at the middle stretch. The BOD concentrations in the Pampa river water after the flood event ranged from 0.2 mg/l to 3.2 mg/l ( Table 2). BOD values in the middle stretches were lower or similar to the previous studies [24,26,27]. However, in lower stretches, particularly at Karuvatta and Kainakary the values were slightly higher. Kainakary is a tourist spot and the wastes from surface runoff during the flood events may have contributed to the high BOD value.
Electrical conductivity (EC) values fluctuated between 30.1 μS/cm and 64.2 μS/cm which were within the permissible limits ( Table 2). The EC exhibits an increased value towards downstream (Table 2). Thomas and Abraham [26] reported that millions of pilgrims taking bath in the Pampa river, especially during the pilgrimage season, besides the agricultural runoff and drainage in the course of Pampa river had led to a high concentration of salt and ions contributing to the higher conductivity value (44.9--87.5 μS/cm). However, after a flood event, the lower value was observed, particularly in the middle stretch (30.1-42.2 μS/cm). In the lower stretch EC was lower than the value reported by Thomas and Abraham [26], but higher than the value reported by Sajudeen et al. [24]. Kainakary station at thelower stretch only showed maximum conductivity (64.2 μS/cm) which may be due to seawater intrusion. Slightly lower values of EC post flood may be due to the dilution of river water by rainwater. Variation in EC may also be attributed to the nature of bedrock of the basin area. Lesser EC values as observed at Ranni, Aranmula and Kadapra stations may be due to the rocky bottom. River underlined by clay minerals exhibits higher EC values as observed in Veeyapuram and Kainakary stations. Variations in the total-dissolved solids (TDS) values resemble those of EC. TDS ranged between 14.9 mg/l (Ranni) to 32.1 mg/l (Kainakary) ( Table 2). Sometimes, high TDS affects the drinking water quality and its high value is associated with high alkalinity or hardness [43]. According to Gaillard et al. [44], TDS content of the majority of the rivers in the world is <500 mg/l and the values of the present study indicate the suitability of water for all purposes. Similar to EC, there was a downstream increment of TDS value which indicates that EC is an excellent indicator for TDS. Compared to a previous study [26] (Table S1), the TDS value in Pampa river was found to be low which can be attributed to dilution by floodwater. Previous reports indicate that tropical small mountainous river contains comparatively low TDS (i.e. <100 mg/l) [45] and the results of the present study also reaffirm the same. However, Thomas et al. [41] reported a high TDS value for tropical mountainous Pambar river in Kerala, India which may be due to semiarid climate and anthropogenic activities.
During the study period, low turbidity value (3.86-4.6 NTU) ( Table 2) was noted in middle (except Aranmula) stretch which was even lower than the reported [24,26] value for the middle stretch of the river Pampa (Table S1). The desirable and permissible limit of turbidity for potability is 5 NTU and 10 NTU, respectively, [31], whereas as per WHO standards [46], it should be ideally below 1 NTU. The post-flood values of turbidity were higher (except Kainakary) at a lower stretch which may be due to the agitation caused by rainfall. It may also be attributed to the deposition of eroded particles from the upper and middle stretches reaches the lower stretch of the river that would have been churned up and re-suspended during further rainfall events.
Total Alkalinity (TA) in the Pampa river system ranged between 19.2 mg/l and 34 mg/l ( Table 2). Though all the water quality indicators are affected by dilution, alkalinity values were higher than the reported value [24,26]. For drinking water, ideally, TA value should be 200 mg/l, while fish can survive at a value 50-300 mg/l [36]. Water with alkalinity <20 mg/l has very low productivity and this study revealed marginal productivity of Pampa river in terms of alkalinity. Comparatively, higher alkalinity values during postflood also supported the higher pH status of the Pampa river water. However, still the values were in the safe limit which indicating water quality of Pampa river was not deteriorated even after the flood. Total hardness (TH) in the water ranged from 8 mg/l to 18 mg/l ( Table 2) and was slightly higher than the pre-flood values as reported by Sajudeen et al. [24] and Thomas and Abraham [26]. However, the values were in conjunction with the reports of Rajan and Anila [27]. The hardness values generally showed an increasing trend towards downstream. Manimala river in Kerala also showed similar spatial variation for TH [40]. A relatively higher value of total hardness at downstream of the river during post-flood period could be due to fertiliser runoff from the Kuttanad rice fields (Thakazhi and Kainakary). For drinking water, recommended TH value is up to 200 mg/l and the optimal value for fish growth is 30-180 mg/l [36]. Kannan [47] has classified freshwater on the basis of hardness as follows: 0-120 as soft, 61-120 as moderately hard, and 120-160 as hard and above 180 as very hard. This study indicates that the Pampa river water is soft even after the flood event. Ca-H ranged from 3.01 mg/l to 7.02 mg/l at various monitoring sites of the Pampa river (Table 2). An increasing trend for Ca-H was noted in the downstream stations. Mg-H (0.12-0.6 mg/l) was lower than the corresponding Ca values ( Table 2) which were also fluctuating from the middle to downstream. The values showed a marked increase towards downstream. Ca and Mg in the Pampa water were slightly lower than the pre-flood values as reported by Sajudeen et al. [24]. Low Ca and Mg content in river water may be attributed to dilution by rainwater. Low values for Ca and Mg were also reported in small mountainous river Meenachil in Kerala [48]. However, values of the present study were much lower as compared to the Muthirapuzha and Pambar river situated in the similar host lithology [49]. Anthropogenic disturbances may be the reason for these differences.
Cl − of Pampa river water ranged from 6 mg/l to 11 mg/l ( Table 2) which was well within the permissible limit of 250 mg/l [50]. Downstream stations showed a higher value for Cl − . Since most of the stations of downstream region fall under Kuttanad rice cultivation area, the mineral fertiliser inputs and high use of water for irrigation would have led to higher evapotranspiration [51], contributing to a higher value [41,49]. Tourism activities in Kainakary and Aranmula stations may have resulted in higher Cl − concentration. Salinity at different stations ranged between 10.8 mg/l and 19.9 mg/l. Compared to the previous study by Rajan and Anila [27], post-flood salinity level increased all along the Pampa river ( Table 2). This contradicts the reports on sudden fall in salinity subsequent to natural calamities like flood or storms [37,52]. However, Dubey et al. [53] reported that flooding can cause sea-level rise which hampers the freshwater aquaculture.  Table 2). In other rivers in Kerala, NO 3 -N has been reported to vary from 0.02 mg/l to 1.30 mg/l in Killiar [54], 4.18-4.83 mg/l in the Nila River [55] which was found to be higher than the present study. All the sites exhibited NO 3 -N level lower than the tolerance limit prescribed for potability. The silicate values ranged from 0.9 mg/l to 4.2 mg/l which is comparatively lower than the Killiar (0.12-5.02 mg/l) [54], Nila (64.83-119.69 mg/l) [55] and Karamana River (1.80-33.6 mg/l) [56]  -P in the present study was lower than the previous report [24] for Pampa water. The low nutrients in Pampa river water could be due to the dilution by floodwater. In the Vembanad lake, forming the estuary of the Pampa, near Kainakary station, low nutrient content was noted which is due to the presence of abundant water hyacinths which absorb the nutrients, that to some extent serve as 'sewage treatment plants'. The water samples of the Pampa river showed lesser spatial variability with reference to nutrients (NO 3 -N and PO 4 3--P) which indicates the influence of flood induced dilution throughout the river stretch.

Heavy metals in Pampa river water
Among the five metal elements viz. Cu, Zn, Cd, Pb and Mn analysed, only Mn could be detected in water samples of Pampa rivers and varied from 0.02 mg/l to 0.095 mg/l (Table  S2). Sajudeen et al. [24] detected only Cu, Zn and Mn in Pampa river and Cd and Pb was in below detection level. Mn detected (0.01-0.09 mg/l) in their study was almost similar to the present study. Mn is a secondary water contaminant and for Mn, USEPA has set secondary drinking water standard with a value of 0.05 mg/l. At some locations (Veeyapuram, Karuvatta and Thakazhi), the concentration of Mn was found to be above the secondary drinking water standard which can lead to black to brown-coloured water and impart a bitter metallic taste to the water [57]. This may cause a great number of people to stop using water from their public water system even though the water is actually safe to drink. In general, flood water decreases the pH of river water resulting in increased solubility of metals and nutrients. Thus, due to higher solubility, metals become more toxic. However, our post-flood survey indicates that Pampa river water became slightly alkaline which may be the reason for nondetection of most of the metals. Dilution may be another reason. However, the frequency of detection and concentration of Mn in the water samples of the Pampa river is higher particularly in the lower stretch as compared to middle stretch. It may be because; the lower stretch comes under the Kuttanad ecosystem where intensive agriculture and aquaculture are polluting the system besides disposal of sewage and industrial effluents as well as unscrupulous use of agrochemicals. The dissolved Mn concentration of the Pampa river was higher as compared with other Indian rivers like Cauvery (2.60 ug/l) [58] and Kali (3.95 ug/l) [59]. The dissolved Mn averages of the Pampa river (22.78 ug/l) was too high than that reported from Kuttanad backwaters (2.41 ug/l), Kerala [60].

Designated use of Pampa river water
For appropriately managing a waterbody, goal for the proposed use of the water has to be defined. According to CPCB [61], India, river water can be classified into five major classes from A to E based on their designated best use criteria ( Table 3). The analysis was also done to find whether Pampa river water is suitable for fish propagation. Results of the present study show that the water quality at all the stations of Pampa river (except Thakazhi and Kainakary) can be designated as 'class A', i.e. use of Pampa river water for drinking purposes without conventional treatment but after disinfection (chlorination) ( Table 4). The water quality of River Pampa was also found to be suitable for fish growth.

Evaluation of Pampa river water post-flood based on water quality index (WQI)
WQI gives a complete image of the aquatic body and potability of the water. For calculating WQI, eleven different quality parameters (pH, DO, EC, TDS, TA, TH, Ca, Mg, Cl − , BOD and NO 3 -N) were selected and their permissible limits as defined by WHO and BIS are given in Table 5 along with their unit weight. Based on the calculated value for WQI, the waterbodies are categorised as follows: Excellent (0-25), Good , Poor (51-75), Very Poor (76-100) and Unsuitable (>100) for potability [62]. Value for WQI ranged from 33.62 to 114.89 for Pampa river and water at various stations can be categorised from 'Good' to 'Unsuitable' category. Ranni and Aranmula station showed the poor quality of water which may be due to rubber plantation and tourism (boat race) activities at respective stations. However, Veeyapuram station where Achenkovil river joins the Pampa river showing 'Good' quality water may be due to the dilution effect. Thakazhi station is an agricultural village entirely with cottage industries and showing 'Good' quality water. The high WQI at Karuvatta and Kainakary station may be due to high BOD and low DO, respectively. Stagnant flow at Kainakary may also be the reason for high WQI as suggested by Vega et al. [33]. Mohan et al. [9] also reported that WQI of Periyar river, Kerala, falls from marginal to poor quality after the August 2018 flood. The overall WQI showed that the river is marginally stressed and a continuous water quality monitoring programme and proper water safety plan are essential to preserve and improve the water quality.

Pearson's correlation study for water quality parameters
The Pearson's correlation matrix is depicted in Table S3. DO has a strong positive correlation with water pH (r = 0.78, p < 0.05). This indicates pH is an important parameter in quantifying the DO content. Dutta et al. [18] reported that high pH reduces the microbial activity and hence the oxygen consumption leading to increase in DO content of the water. In the present study, DO shows a significant negative correlation with EC, TDS, TH, Ca-H and Mn, which may be due to the addition of different biodegradable pollutants from domestic sewage, municipality's wastes, run-off from agricultural land, etc. washed into the river surging the BOD. Thus, the concentration of DO gradually depletes and higher values of other water quality parameters. The present observation is in agreement with findings by Barat and Jha [63] and Chakraborty [64]. Further, there was a nonsignificant inverse relationship between DO and temperature, which is quite obvious. Naturally, transparency is inversely proportional to turbidity and the present study reveals a significant negative relationship between them. Inflows of turbid rainwater into the river water may have reduced the transparency. TDS showed a significant positive correlation with TH and nonsignificant positive correlation with Ca-H and Cl − . TH also showed a strong positive correlation with Ca-H and weak positive correlation with Cl − , indicating a predominant form of Ca and Mg as chloride salt. Temperature is positively correlated with EC and TDS. PO 4 3-P was strongly correlated with Ca-H. The positive association between these components can be explained due to homogeneity in their distribution pattern or they have a common origin. Mn had an inverse correlation with pH, DO and NO 3 -N. Mn is a nutrient-type metal and its enrichment may contribute to the significant negative correlations between Mn and pH, DO and nutrients [65].

Principal component analysis
Based on the eigenvalues, principal components were identified for understanding basic data structure [66]. Scree plot of eigenvalue indicates that after the fourth component, there is a significant change in the slope of the plot and thus eigenvalues above unity were considered which explains 93.8% of the variance in the dataset [67].
Loading of PCs (Table S4) helps in understanding the basic nature of a component corresponding to the correlation involving PCs and original variables [33]. There are several criteria to decide the significant variables and in this study, we have considered variables with factor loadings >0.5 as a major contributing factor. PC1 explains 45.01% of the variance and contributed significantly by depth, temperature, EC, TDS, TA, TH, Ca-H, Cl − , salinity, PO 4 3 -P, and Mn. Correlation matrix also showed the correlation of these Due to the highly random nature of PCs, it may interfere with the interpretation of water quality parameters. To get a more pertinent and simpler interpretation of hydrochemistry of Pampa water, PCs were rotated (varimax) to get variables which are more significant by using factor analysis (FA). Factor loadings of the varimax rotated components (called varifactors) are presented in Table S5. The total variance of 36.65% was  explained by varifactor 1 and taken part by depth, temperature, EC, TDS, TA, TH, and Ca-H,  PO 4 3-P and Mn which can be coded as the contribution of a mineral segment of river water. Varifactor 2 explained 20.7% of the variance and is mainly contributed by Cl − , Salinity, BOD and SiO 2 which can be interpreted as the high salinity conditions with high silicates that may lead to increase in phytoplankton growth. Varifactor 3 (20.05% of the variance) was positively contributed by depth and transparency and negative by turbidity and Mn. This indicates the positive correlation between depth and transparency and negative relation between turbidity and transparency. This also explains that the stations with higher depth are more transparent. Varifactor 4 (16.45% of the variance) is positively contributed by Mn and Mg-H whereas pH and NO 3 − contributed negatively which may be due to the polluted water washed along with the floodwater into the river. Both strong positive and negative loadings in factorial analysis indicate that some attention is required in rejuvenating the health of Pampa river [68].

Cluster analysis
Cluster analysis (CA) is a way to represent a big dataset into small clusters or groups based on their characters. Clusters which were formed after an analysis showed high homogeneity (within the groups or clusters) or heterogeneity (within groups or clusters). Results of cluster analysis are usually demonstrated by a dendrogram (tree diagram) which summarises the visual demonstration of the clustering process based on propinquity or remoteness of the dataset. Here, datasets were clustered by employing Ward's linkage and Euclidean distance method. The dendrogram ( Figure 2) reveals that upstream stations are significantly different from downstream stations. However, Veeyapuram being a downstream station is clustered with the upstream station. Pampa river joins Achenkovil river at Veeyapuram station and the inter-basin transfer might have modified the geochemistry of the water at this station causing a mixed nature. Aranmula station is clubbed together with the downstream station which may be due to higher turbidity, BOD and chlorinity as compared to other middle stream stations. Kainakary is distinctly different from other downstream stations. Pampa joins Vembanad lake at the tip of Kainakary and lake process might have played an important role in clustering this station differently. These findings point out towards spatial differences that played an important role in determining the water quality of the Pampa river.

Conclusion
The magnitude of the massive flood and the nature of the affected catchment area made it essential to assess the quality of Pampa river water to know the suitability of the riverine system for different water uses including fisheries. Such studies are useful for disaster management in the event of unprecedented natural calamities. Due to the erratic behaviour of floods (enormity and rate of recurrence), baseline data for water quality immediately before the flood is absent which could have given an insight into the environmental processes in accounting the quality of water. However, a comparison was made with available previous reports. Hence, the present study may serve as baseline information on Pampa river water quality after the flood event and the following conclusions may be drawn: • Post-flood the water quality parameters of river Pampa showed transitory effects, which were more pronounced along the lower stretches. • The reduced concentration of water nutrients and metal content attributes to the dilution effect due to the flood. It seems that despite the devastating effect of August 2018 flood, the environment has benefitted by a reduction of metals in Pampa river water at least for a while. • The pre-flood neutral to the acidic property of water shifted towards alkaline conditions throughout the river stretch post flood. • Alkalinity and salinity were almost twice higher than the pre-flood scenario, though the values were within the ranges required for fish survival and growth. • Decreased conductivity and increased dissolved oxygen are also the major changes observed post flood. • Water quality of Pampa river falls in between 'Good' to 'Unsuitable' category which to some extent is in an awful state and requires immediate rejuvenation action particularly in downstream stations where anthropogenic activities are at a higher rate. • Multivariate statistical techniques like PCA and CA were effectively applied for evaluation of geospatial variation in the Pampa river water quality which identified the important factors controlling the water quality after the flood.

Recommendation
The study was carried out for a short period and a single season, immediately after the August 2018 Kerala flood. Though this study could shed some light on the impact of flood, repeated studies in time and space in the future are required to assess its real impact.
Sediment plays a crucial role in determining the overlying water quality and heavy precipitation may change the sediment texture and composition which in turn define the water quality as well as the benthic community structure. Thus, future study should also include the sediment quality analysis and influencing mechanisms of these sediment properties on the distributions and fates of nutrients and heavy metals in water. Flood induced changes in water quality can have direct impacts on phytoplankton and zooplankton communities and ultimately on fisheries. Therefore, future endeavours should cover the plankton and fish diversity studies. This work should also be extended to modelling studies which will help to predict the status of the riverine ecosystem for developing suitable management practices to ensure sustainability.