Waterbird abundance and diversity along a stretch of Ganga River, India modified by construction of a barrage

Damming and diverting river water alters the channel characteristics and natural flow regime. The change in biotic and abiotic factors results in dissimilar habitat conditions upstream and downstream of the barrage. Given the habitat dissimilarity and therefore resource availability, we hypothesized the dissimilarity in waterbird abundance and species diversity in the river habitat upstream and downstream of the barrage. The study was conducted on a 24 km stretch of Ganga River at Narora, Uttar Pradesh, India. This stretch overlaps with a Ramsar site as well as an Important Bird and Biodiversity Area (IBA). Bird sampling was done using transect count method for summer and winter season in 2017. The abundance was studied in relation to three habitat variables viz. channel depth, channel width and anthropogenic disturbances using Generalized Linear Model. A total of 140 species of birds were recorded. There was statistically significant difference in the abundance of waterbirds between upstream and downstream stretch in winter season (Mann-Whitney U, p < 0.05). During winters, migratory waterbirds especially ducks and geese occupied the deep water upstream of barrage, while the downstream was mostly occupied by terns and cormorants. Waterbird species diversity was higher in downstream during winters (Effective Number of Species=28; Shannon’s Index= 3.35) and higher in upstream during summer (Effective Number of Species=25, Shannon’s Index=2.87). Channel width, channel depth and anthropogenic disturbance influenced waterbird abundance in both the seasons (Generalized Linear Model; p < 0.05). The influence of channel depth varied with season for the upstream stretch.


Introduction 57
Our dependence on the surface freshwater sources is evident from the fact that "over 50% of the 58 world's population lives closer than 3 km to a surface freshwater body, and only 10% of the population 59 lives further than 10 km away" [1]. Rivers around the world are being heavily modified by dams, weirs 60 and barrages to cater for irrigation, electricity generation and drinking water requirements. These

79
The barrage has created reservoir like conditions in the upstream with a greater water depth and partly 80 confined channel due to hard and muddy banks. In contrast, the downstream stretch of river has highly 81 braided and unconfined channel and more sand bars and point bars. The vegetation along the bank is sparse. Part of the river stretch being a Ramsar Site, regulates heavy fishing activities, but the 83 unregulated stretch is disturbed with fishing nets.

84
Birds are conspicuous and inhabit wide variety of habitats. Changes in their population density, 85 abundance or distribution often indicate changes in environmental conditions [12][13][14]. These changes 86 can be the modification of their habitat, scarce food resources or absence of safe nesting sites. However, 87 higher abundance or density of birds in a habitat does not indicate its higher quality [15]. Habitat 88 selection is a behavioural response of selection of a set of specific physical environmental conditions 89 that influence the survival of the individuals of a species [16]. Nonetheless

106
Given the close association between bird morphology and habitat selection and observed 107 dissimilarity in habitat in the river stretch upstream of barrage (henceforth USB) and downstream of 108 barrage (henceforth DSB), we hypothesized that in the available habitat, presence of birds (habitat use) in either stretches is governed by their food habits and body size and therefore the species diversity 110 and abundance of waterbirds is different in the two stretches. We also investigated the significance of 111 habitat characteristics which have been modified due to construction of barrage viz., channel depth and 112 channel width and the effect of anthropogenic disturbance in influencing waterbird abundance.

152
The bird sampling was done in the month of February 2017 for winter season and in the month of 153 May 2017 for summer season. Five replications of each transect were made in both the seasons, making 154 total field effort equal to 242 km for the study. Birds were continually recorded from the river bank while 155 walking the transect on foot at a speed of 1 to 2 km/h on the right bank of the river. Birds sighted in the 156 main channel, side channels, water pools, islands, sand bars and a strip of 10 m on the bank were 157 recorded. The channels bifurcated only on the right bank of the river were surveyed. The birds were 158 recorded during early morning (30 minutes after sunrise to 11 am) which corresponds to high bird 159 activity. Birds were identified based on sightings only. For each sighting, GPS location (using hand-held 160 GPS-Garmin 12XL), time and habitat was recorded. Only birds flying from the opposite direction parallel 161 to river were counted and those flying across the river were excluded to avoid double counting.
162 Observations were made for all the birds but only the count of waterbirds and water-associated birds 163 were used in analysis. Opportunistic sightings were included in the checklist but not used in the analysis.
164 Field Guide by [37] was used for identification. Based on food habits, food acquiring strategies and body 165 size, we grouped the waterbirds and water-associated bird species into seven guilds (S2 Table 1). 173 inclinometer accuracy= ± 0.1 degree) at a distance of every 100 m for the entire stretch. For analysis, 174 mean depth and mean width were computed for each 1-km stretch of the river (Fig 2). Fourteen major 175 categories of anthropogenic disturbances were identified in the study area based on observations from 176 all ten complete surveys (S1 Table). Each category was scored on a scale of 0 to 3 depending upon 177 intensity of disturbance for waterbirds. The scores 0 to 3 signify absence, low disturbance, moderate 178 disturbance and high disturbance in the order. Low and moderate categories were assigned relative to 179 the highest intensity of disturbance observed in the entire stretch and based on the observations made 180 during all the ten visits. For each one kilometer section of the river, the scores corresponding to each 181 category were added. Therefore, minimum possible score is zero and maximum possible score is 42. The bird count for each species from five replicates was used to compute maximum count 188 (mode) for that species for a particular season. Maximum count was also computed for each guild 189 considered in this study. Simpson

200
Overall diversity 201 A total of 140 species of birds were recorded within a 24 km stretch (S2 Table). These species

207
Majority of the birds were carnivore (62 species) or omnivore (61 species) (S2 Table) Table).  (Table 1). Overall abundance declined as winter 231 migrants left the site. It was observed that number of ducks and geese declined when season changed 232 from winter to summer but abundance of egrets, storks and ibises increased (Fig 3). Additionally, in 233 winter, the percentage count of diving ducks was higher in the USB and in summer, it was higher in the 234 DSB. Waders and dabbling ducks were more abundant in the DSB during summers as compared to 235 upstream (Fig 3).       (Table 3). (2) Downstream-Winter Analysis of the influence of habitat variables for different guilds revealed that channel depth and 264 width are significant predictors of abundance for all the guilds (Table 4). Disturbance had 265 statistically significant influence on abundance of all the guilds except for G5 (Table 4). Stretch

266
(USB or DSB) was found to be significant for all guilds except for G7 indicating that the abundance 267 of water-associated birds is not influenced by the stretch type. Season did not influence the 268 abundance of guilds G3, G6 and G7 (Table 4).

286
During winters, river depth is the positive influencing factor in the USB. This positive influence of depth 287 in USB was due to presence of reservoir and increased surface area of water which is preferred by 288 diving ducks and dabbling birds [32]. It was observed to be a preferred habitat by Common Pochard,

289
which was only observed in USB. In summer, we observed a negative influence of water depth on 290 waterbird abundance. This observation can be attributed to change in species composition with change 291 in season, where diving and dabbling ducks were dominating in the winters but waders, herons and 292 kingfishers were more abundant in the summer season and occupied shallow water habitats (Fig 2). In 293 DSB, where water depth is lower than USB, a mosaic of habitat due to exposed point bars and islands is 294 available which provided additional habitat for shallow water preferring birds [45]. Waders like small 295 pratincole and little ringed plover were exclusively found in the downstream stretch, as they prefer 296 areas with lowest vegetation cover [46].

297
Similarly, we found that channel width positively influences the abundance of diving ducks (G1)

318
This study was a preliminary attempt to understand the difference in waterbird abundance and 319 diversity in the partially lacustrine and lotic habitat conditions created by a water diverting structure 320 on Ganga River. Also this study is the first intensive study that covers part of both the Upper Ganga