Seasonal dynamics of ichthyodiversity in a hill stream of the Darjeeling Himalaya, West Bengal, india

Author Details: M.L. Acharjee, PhD in zoology from University of North Bengal, Assistant Professor and Head, Department of Zoology, Kalimpong College. Published six research paper and several articles. S. Barat, PhD in zoology, he is Professor and Head of the Department of Zoology, University of North Bengal. He runs Aquaculture & Limnology Research Unit in the University and published more than 45 research paper in the reputed journals.

The Darjeeling Himalaya has a pronounced seasonal climate and lies north of the tropical belt between 27 0 13'05"-26 0 17'10"N and 88 0 53'00"-87 0 59'30"E. The spring-fed torrential hill streams of the region represent a unique lotic water ecosystem characterized by rocky and gravely bottoms, high transparency, low temperature, high oxygen level and water velocity, complex flood regimes and seasonal variations in volume of flow (Acharjee & Barat 2011). Sub-tropical regions with monsoon climates experience some of the largest floods on earth, coupled with highly predictable seasonality (Gupta 1987). The Darjeeling Himalaya has five climatic seasons, with the winter and monsoon seasons being the longest. The impact of fishing methods and gear on overall fish diversity (Acharjee & Barat 2010) and diversity of loach and catfishes (Acharjee & Barat 2012) have been studied in the Relli hill-stream, and the ichthyofaunal diversity of the Teesta River in this region (Acharjee & Barat 2013) has also been examined. But to date there have been no studies of seasonal dynamics of density and diversity indices of ichthyofauna in the area. The present study examines the seasonal dynamics of distribution and abundance of fish species in different reaches of the Relli hill-stream. These observations provide insights into seasonal changes and disturbances of the stream ecosystem, which may be helpful in the conservation and management of fish species and formulating fishery policy.

Study area
Relli or Rilli is a spring-fed torrential tributary of the River Teesta (Rilli is the Lepcha word for swirling river). The stream originates in the Algarah-Lava ridge of Kalimpong subdivision in the Khampang Reserve Forest at an altitude of about 1800m. The total length of the Relli stream is about 32km (entirely within Kalimpong sub-division) and it joins the Teesta at an altitude of about 212m (Fig. 1). The average width of the stream bed varies from 90-150 m. The longitudinal slope of the stream is about 11m.km -1 . The slope of the stream bank varies from 20-90 0 . During monsoon the stream is quite swift but during dry seasons it is greatly reduced. The catchment area of the stream is 165km² and is hilly and well managed. The stream flowing through a gorge is flanked by two ridges. One ridge coming up from Sepkhola-Kamesi forest (300-800 m) traverses across Durpin (1550m), Kalimpong Town (1250m) on its saddle. Then it rises up to Delo (1800m), to Algarah (1900m) and to Labha (2150m) The other ridge arising from Mungpong-Pankhaban forest (250-400 m) leads to Yangmakum (800-1200 m) and further to Charkholey (1600m) from where it meets with its sister ridge arising from Bagrakot (250m) across Nimbong (1000-2000 m).
In this study three sampling stations (RR-I, RR-II and RR-III) were selected for the collection of fish. Physical characteristics and channel morphology ( Table 1) and descriptions of sampling sites are as follows: RR-I is located at a distance of 10km from the origin and near the confluence with Payong Khola, this site has little disturbance aside from fishing. RR-II is located 6km from RR-1 near the confluence with Pala Khola, this site is subject to activities like sand and boulder mining, sport and subsistence fishing and car/clothes washing and is a popular picnic spot. RR-III is located 8km from RR-2 near the confluence with Khoni Khola, it is also relatively undisturbed with little fishing activity.

Climate and Seasons
The configuration of neighbouring mountains plays a key role in shaping of the climate of Darjeeling Himalaya. The region has five climatic seasons namely, (i) spring, (ii) summer, (iii) monsoon/rainy, (iv) autumn and (v) winter. These seasons are usually marked from the mid of the months following the English calendar. Spring usually prevails from the mid February to mid April, when the region experiences frequent hail storms and occasional showers. Summer starts mid-April and continues to mid-June with hot weather. The rainy season starts mid-June and lasts until mid-September, during which the district experiences continuous rainfall, thick clouds, fog, mist and high relative humidity. Autumn starts in the latter half of September and lasts until November when winter begins, lasting to mid-February; temperatures can drop below freezing in mountainous regions above 2600m. The lower hill region below 1500m does not receive snowfall, but occasional accumulation of frost is observed. For convenience, the seasons of the study area expressed as pre-monsoon (March, April, May), monsoon (June, July, August), postmonsoon (September, October, November) and winter (December, January, February).
The annual precipitation observed in the first year of study (March 2007to February 2008 was 2394mm (with 103 rainy days), and in the second year (March 2008 to February 2009) it was 1895 mm (with 89 rainy days). Maximum rainfall occurs during the monsoon months; in the first study year it was 2157mm with 74 rainy days (highest in July 962mm with 29 rainy days), and in second year it was 1615mm with 65 rainy days (highest in August 649mm with 17 rainy days). High rainfall in the entire hilly catchments of the hill stream during monsoon months results in some flooding.

Fish sampling
Monthly sampling was carried out (from March 2007 to February 2009) in the stream in three sampling sites following fishermen or local people who catch fish in this region.
-About 1000m 2 of stream channel in each site was blocked by nets and traps (closed population) to prevent movement of fishes in and out of the site.
-Using cast net (Bhureli jal) which is small in size (r =1.2m; weigh about 5kg) having a mesh ranging from 0.5-1.5 cm with an average 1.0cm and covering an area of about 4.5m 2 , three separate and sequential efforts were made within the blocked stream. Effort (one person, casting nets 25 times, covering an area about 100m 2 ) was kept constant. Following each efforts, captured fish were put into separate bucket with effort number. The catch for efforts 1, 2, 3 was denoted as n 1 , n 2 and n 3 , respectively.
-In the shallow water rock-striking method was also used. Area of a flat stone was measured for this purpose on which a weighty iron hammer is struck with full force. The intensity, vibration and sound thus produced causes fish to float over water-surface which were hiding under the stone. The fishes were scooped with nets and density calculated and compared with the catches of cast net.
-Other gear and methods including scoop nets, hook and line, various traps (Dhadiya, tip, thokre and thali trap), electro-fishing and water diversion were employed when available. The only species taken into account were those not found in the catches of cast nets.
-Removal or depletion estimates (Li & Li 2007) were made by extrapolation of CPUE (y-axis) vs. accumulated catches (x-axis). The population estimate N also calculated using simple linear regression and determining the x-intercept for the cumulative catch. It was expressed as estimated population size of ichthyofauna (N) per 1000m 2 of river channel . CPUE = intercept y -slope x ∑Ci and, when CPUE = 0, N = intercept y /slope.
-The fish caught were examined for morphological features, colour bands or spots present on the body, recorded in the field along with length and weight. Photographs were taken by digital compact camera (Nikon, coolpix L10). Gut contents were collected for analysis of food habits. Two specimens of each species were taken for identification and their species were ascertained on the basis of morphometric characters and meristic counts following the criteria given by Shaw & Shebbeare (1937), Talwar & Jhingran (1991), Sen (1992), Tekriwal & Rao (1999), Nath & Dey (2000, and Jayaram (2009). Valid scientific names were taken from Fish Base (Froese & Pauly 2012).

Diversity Indices
The following diversity indices were calculated to understand icthyodiversity.
Where, S = total number of species; N = total number of individuals.

Shannon-Wiener Diversity index (Krebs 1999).
Where, H' = Information content of sample (bits/ individual) = Index of species diversity; S = Number of species; pi = Proportion of total sample belonging to ith species (ni/N); ni = number of individuals of i th species in the sample; N = Total number of individuals in the sample = Σni.

Statistical analysis
To test for seasonal differences (four seasons namely pre-monsoon, monsoon, post-monsoon and winter) of density and diversity indices of ichthyofauna, one-way ANOVAs and Fisher Least Significant Differences (LSD, a post hoc test) were performed. To account for multiple comparisons Bonferroni correction was applied (using SPSS vers. 16 software).

RESULTS AND DISCUSSION
A total of 25 species of fish were recorded from the hill-stream belonging to five families and 16 genera (Tables 2 & 3). The largest number of species (14) were from the family Cyprinidae, followed by Balitoridae (S=6), Sisoridae (S=3), Psilorhynchidae and Cobitidae (one species each). Devario aequipinnatus, Crossocheilus latius, Cyprinion semiplotum, Psilorhynchus sucatio, Acanthocobitis botia, Somileptes gongota, Euchiloglanis hodgarti and Glyptothorax cavia were found only in site RR-III. Barilius shacra and Garra lamta were found in sites RR-II and III, and Schizothorax progastus was found in sites RR-I and II. Twenty-two species (88%) belong to order Cypriniformes and three (12%) to Siluriformes. Nelson (1994) reported the greatest diversity in Cypriniformes and Siluriformes in freshwater habitats, and similar observations were made by Johal & Rawal (2005) in a western Himalayan hill stream, and by Shrestha (1999) in the Nepal region of the Himalayas.
Seven species of fish are abundant (78.6%) in Relli: Neolissochilus hexagonolepis (15.8%), Garra gotyla (14.5%), Barilius bendelisis (14.3%), Schizothorax richardsonii (10.1%), Schistura devdevi (9.8%), Opsarius barna (8.3%), and Schistura scaturigina (5.8%). Analyses of gut contents revealed that Neolissochilus hexagonolepis, Barilius bendelisis and Opsarius barna are omnivorous, feeding on diatoms, filamentous algae, insect larvae etc. Schistura devdevi and S. scaturigina are carni-omnivorous and feed upon insect larvae, crustaceans and algae, whereas Garra gotyla and Schizothorax richardsonii feed upon algae and detritus, and hence are herbivorous. Thus the analysis of the trophic structure indicates dominance of omnivorous fish in all sites, followed by carni-omnivorous and herbivorous fish. Similar findings were reported from Yangtze River basin of China (Fu et al. 2003) and two tropical rivers of India (Das & Chakrabarty 2007). Based on the study of trophic levels it appears that omnivores are the most tolerant of degradation or ecosystem dysfunction because they are able to consume food from a wide variety of sources in a changing ecosystem (Wichert & Rapport 1998). This may be useful to assess the quality of fish habitat.
Almost all species in the collection are represented by very young specimens, adult brood fishes were found mostly during monsoon and post-monsoon months. The array of juveniles and sub adults in this work suggests this stream is probably used as a breeding ground and nursery ground for fish species, particularly of Schizothorax richardsonii, Schizothorax progastus, Neolissochilus hexagonolepis, Neolissochilus hexastichus. Hora & Gupta (1940) found very young fish specimens in Kalimpong duars and Siliguri Terai, and opined that the fish species of this region breed during and after the monsoon months. Hora & Nair (1944) reported mature fish with developed gonads in Riyang stream, a small tributary of the river Teesta in Darjeeling Himalaya, and also found small fry of some species in the same stream. They opined that the Teesta provides a highway for fish migration. However, the absence of adults, among other factors, may be due to pressures exerted from overfishing by local fishermen and/or from downstream migration of fish. Similar observations have been made by Emmanuel & Modupe (2010) in the rivers of Nigeria. Acharjee & Barat (2010) reported that over the years, uncontrolled and often indiscriminate fishing in the largely unmanaged Relli stream has resulted in a sharp decline in catches of important sport and subsistence fish.
It is evident from the present investigation that the ichthyofaunal composition of the 14km stretch (altitudinal variation from 704-340 m) was represented mostly by rheophilic, cold water hill stream species similar to cold water species of Nepal, Sikkim, Bhutan, Assam and other parts of the Indian Himalaya. Dominance of rheophilic, cold water hill stream species may be due to the nature of the stream, which has high water current (ranged from 0.61m sec -1 to 1.51m sec -1 with a mean value of 0.90m sec -1 ±0.24) and gradient, and low water temperature (ranged from 13.3 0 C to 27.8 0 C with amplitude of 14.5 0 C and mean value 21.4 0 C ±3.4) (Acharjee & Barat 2011). From the fishery point of view, water temperature falling with the tolerance limit of trout may be termed cold (Jhingran & Sehgal 1978 Table 3. Occurrence, distribution, abundance, local name and other status of fish species at the sampling sites of Relli River during the study period.  (Hora, 1923) L=6; Torrent Catfish (E); Til-kabri(N).

Ichthyodiversity in a Darjeeling hill stream Acharjee & Bharat
The fast flow of the Relli places constraints on its fish fauna, and the species observed in the present study can be categorised into six major groups on the basis of morphological characteristics enabling them to inhabit torrential streams. These groups are: (a) fish living in shallow clear cold waters in the foothills without obvious modifications to current, such as Barilius spp.; (b) fish inhabiting the bottom water layers in deep fast current with powerful muscular cylindrical bodies, such as Schizothorax spp., Neolissochilus spp.; (c) fish sheltering among pebbles and stones to avoid the strong current, including Crossocheilus latius; (d) fish sheltering among pebbles and shingles in shallows with special attachment structures, such as the loaches, Schistura spp., Canthophrys gongota, Aborichthys elongatus, Acanthocobitis botia; (e) fish that cling to exposed surface of bare rocks in slower current with adhesive organs on their ventral surfaces, such as Garra spp., Glyptothorax cavia, Pseudecheneis sulcata; and (f) fish that cling to the exposed surfaces of bare rocks in fast current having limpet-shaped bodies and mouth, gills and fins highly modified to suit their habitat, such as Parachiloglanis hodgarti. Similar groupings have been recognised by Menon (1954) in Himalayan fish.
The stream has pool-run-riffle geomorphology and high substrate diversity (Bedrock, cobble, gravel, sand and terrestrial litter), and is subjected to seasonal discharge variation including periodic spates and flash-floods during the wet season. Dudgeon (2008) mentioned that high elevation streams in the tropics have fast currents, and are characterized by plunge pools, riffles and rapids dominated by boulders and cobbles, with few or no aquatic macrophytes. The present study revealed that almost all the cyprinid fishes especially Barilius spp, Devario aequipinnatus are generally distributed in the pools, pool edges and in the side pools. Small fry of various fish species such as Barilius spp., Garra spp., Neolissochilus spp., Tor spp., Schizothorax spp., Nemacheilus spp., etc. were found in the side pools. Schistura spp., Aborichthys elongatus were generally abundant in riffle type microhabitat and found under stones.
Of the 25 species, 17 species are considered as food fish, 21 species are ornamental fish, four species namely Schizothorax progastus, S. richardsonii, Neolissocheilus hexagonolepis, N. hexastichus are considered sport fish and are commercially important. References to conservation status of fish species and categories within this paper are based on IUCN (IUCN, 2013) classification. Out of the total species recorded, 20 are Least Concern (LC), two are Vulnerable (VU) and three are Near Threatened (NT). This categorization may be useful for planning conservation strategies for genetic resources of ichthyofauna.

Population size and number of species
A summary of ichthyofaunal density and diversity observed in the Relli is presented in Table 4. Analyses of monthly data indicated that CPUE, fish population size (N) and number of species (S) increased slightly during April-May and increased sharply during October-November (Fig. 2). Catch per unit effort (CPUE= one person, 25 casting a net covering an area about 100m 2 ) ranged from 11-53 with a mean value 28 (±9.68). Lowest . Temperature slowly increased through mid February to mid April and the region experience frequent hail storm (spring season). Based on one way ANOVA, CPUE, estimated population size (N) and number of species (S) showed significant seasonal differences and the results of LSD (post hoc tests) with Bonferroni correction revealed that estimated seasonal means were also significant (Figs. 3,4 and 5). Increased water temperature during March, April and May (pre monsoon), availability of fish food (macro invertebrates, plankton) may have a role in the slight upraised population and species richness. The Rainy season experiences continuous rainfall, including periodic spates and flash-floods. Increases in discharge during the monsoon (June, July and August) coincided with increased movement and upstream spawning migrations of many adult fishes, such as Schizothorax richardsonii, S. progastus, Neolissochilus hexagonolepis, N. hexastichus, Garra spp., Cyprinion semiplotum etc. Similar spawning migrations occur with increased discharge of rivers in Nepal (Edds 1993) Africa and South America (Lowe-McConnell 1975;Goulding 1980). Hora & Gupta (1940) opined that the fish species of this region breed during and after the monsoon. Spawning and breeding of Katli (Neolissochilus hexagonolepis) season lasts from June to September (Barat et al. 2005). Due to breeding migration and availability of basic fish food items a good variety of fishes attracted to hill stream (Bahuguna & Badoni 2002). A similar opinion has also been given by Singh & Kumar (2000). Bisht et al. (2009) recorded maximum fish fauna during the monsoon season in a small hill-stream Dangchaura (Takoli) Gad along with River Alaknanda, and this was correlated with the migratory behaviour Months ( of some fishes and changes in the physico-chemical nature of river water. But in the present investigation, monsoonal fish yield was low in the hill stream Relli. It may be due to further migration of fishes to tributary streams like Pala, Payong, Khani khola etc. with clear oxygenated water than the Relli during monsoon or it may be due to difficulty in catching fish with cast net in torrential elevated water. Begossi et al. (1999) opined that during monsoon fish yields decline due to the elevated water level. During the post monsoon (September, October and November) temperature decreases slowly and migratory brood fishes along with year old young begin their downward