Fish (Spinibarbus Hollandi) Dynamics in Relation to Changing Hydrological Conditions: Physical Modelling, Individual-Based Numerical Modelling, and Case Study
Total Page:16
File Type:pdf, Size:1020Kb
ECOHYDROLOGY Ecohydrol. 6, 586–597 (2013) Published online 16 April 2013 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/eco.1388 Fish (Spinibarbus hollandi) dynamics in relation to changing hydrological conditions: physical modelling, individual-based numerical modelling, and case study Rui Han,1 Qiuwen Chen,1,2* Koen Blanckaert,1,3 Weiming Li2 and Ruonan Li1 1 RCEES Chinese Academy of Sciences, Beijing, 100085, China 2 China Three Gorges University, Yichang, 443002, China 3 Department of Limnology of Shallow Lakes and Lowland Rivers, Leibniz-Institute of Freshwater Ecology and Inland Fisheries (IGB), Berlin, Germany ABSTRACT The paper reports the development of an individual-based fish dynamics model, of which the key components are the rules for the movement of individual fish and the definition of the habitat suitability. The distribution of the fish mainly depends on the flow conditions (velocity, depth, substrate) and life cycle of the fish. A major contribution is the refinement of the rules for fish movement, based on laboratory experiments under volitional swimming conditions, which also provided the ranges of preferential velocities and substrate size for the target fish, Spinibarbus hollandi. Moreover, they provided data on the fish trajectories and distribution patterns that allowed for validation of the movement rules. The validated fish dynamics model was applied to investigate the effect of discharge increase during the dry season by means of reservoir operation in the Lijiang River, which was the subject of field investigations in 2007 and 2008. The model results indicated that reservoir operation leads to an increase of fish biomass. According to the fish movement rules, fish cannot always escape from riverbed regions that dry during decreasing discharge events, which causes them to be trapped and die. Reservoir operation decreases the area of dry riverbed and reduces the travel distance for fish to escape from dry regions. Critical advantages of the individual fish model over global models defined on the population level are that they can account for the time that the fish needs to reach a region of suitable habitat and for the spatial pattern of suitable zones and their connectivity. Copyright © 2013 John Wiley & Sons, Ltd. KEY WORDS fish behaviour; fish dynamics; laboratory scaled model; numerical simulation; individual-based model; vector- based movement Received 1 February 2012; Revised 16 February 2013; Accepted 14 March 2013 INTRODUCTION dynamics in relation to physical and environmental parameters (substrate, flow depth, flow velocity, food Fish are an essential component in the river ecosystem and may resources, dissolved oxygen, etc.) predicted by a hydro- be an important economic resource. Therefore, predictions of logical model. Most models do not represent individual fish changes in the fish population in response to natural or behaviour but use global aggregated parameters such as anthropogenic causes, such as climate change, floods, population abundance or biomass. Such models parameterize droughts or hydrological changes induced by reservoirs are individual interactions in a simplified way (Ekeberg 1993, important (Brookshier et al., 1995). The total biomass, spatial Steel et al., 2001, Li et al., 2011a, 2011b). These global distribution and dynamics of the fish are determined by the models merely focus on equilibrium conditions (Crowder fish behaviour (Railsback et al., 1999), which depends on et al., 1992) and are not appropriate for investigating the multiple interrelated physical, physiological and environ- fish dynamics in response to changes in physical or mental processes, as well as interactions between individuals. environmental parameters. These global models cannot, Recently, the progress in computational capacity and forexample,representwhetherornotfish is able to escape spatial data collection (Chen et al., 2010) has facilitated the to shelters during important flood events or avoid being development of a variety of numerical models for fish trapped on dry riverbed during droughts. Moreover, dynamics (Rose et al., 1996). These models predict the fish validation of global models is complicated by the difficulty to measure experimentally the parameters, which are defined at the population level. Most species have mainly been investigated at the individual level (Reed, 1983), *Correspondence to: Qiuwen Chen, RCEES Chinese Academy of Sciences, which favours the use of individual-based models that Beijing, 100085, China. E-mail: [email protected] directly parameterize the behaviour of individual fish Copyright © 2013 John Wiley & Sons, Ltd. FISH DYNAMICS IN RELATION TO CHANGING HYDROLOGICAL CONDITIONS 587 (DeAngelis and Cushman, 1990; Bian, 2003; Humston et al., 3. To apply the validated model for the prediction of the 2004; Li et al., 2010). Calibration and validation of such effect of flow regulation through reservoir operation on models, however, critically rely on experimental data on the Lijiang River in China, in which Spinibarbus individual fish behaviour (DeAngelis and Gross, 1992). At hollandi is the dominant fish species. present, there is a paucity of such experimental data. Experimental studies on fish behaviour, including SPINIBARBUS HOLLANDI swimming orientation and speed, flow cues for migration, preferences for velocity and flow depth, have attracted Spinibarbus hollandi (Osteichthyes, Cypriniformes, increasing interest (Coutant, 2000). Bailey and Batty Cyprinidae, Spinibarbus) is largely present in southern (1983) examined predating behaviour by Aurelia aurita Asia. It is an important aquaculture species in South on early first-feeding stage larvae of the herring Clupea Chinese rivers, such as the Yangtze River, the Pearl River harengus. Webb (2002) investigated the posture, depth and and the Lijiang River. swimming trajectories of various fishes under different Spinibarbus hollandi (S. hollandi) (Figure 1) is charac- conditions of flow perturbation and turbulence. Burrows terized by a blunt and protractile snout, small eyes on the (2001) and Gibson et al. (2002) studied juvenile plaice upper side of the head, a slightly oblique mouth, a posterior Pleuronectes platessa behaviour in relation to depth end of the upper the jaw that reaches the anterior margin of changes. Bégout Anras and Lagardère (2004) investigated the eyes, two pairs of barbels, with the maxillary barbels swimming behaviour of rainbow trout as a function of longer than the mandibular barbels at the corner of the stocking density. Pavlov et al. (1994) and Skorobogatov mouth, an elongated cylindrical body, large and cycloid et al. (1996) investigated effects of turbulence on fish scales, and a complete lateral line. Its dorsal fin origin is in behaviour. These investigations were all performed in front of its pelvic fin origin, its pectoral fin ends are distant laboratory flumes or tanks of simple geometry, which ignore from its pelvic fin origin, its pectoral and pelvic fins are the heterogeneity in physical and environmental conditions situated at the lower side of its body, and it has a forked found in natural rivers. Moreover, the fish behaviour in these caudal fin. experiments was investigated under conditions of forced The life cycle of S. hollandi can be roughly divided into six swimming, which are not representative of conditions stages (Yin, 1998): embryo (1–2 days), larva (3–4days), encountered in natural rivers. Only recently, fish behaviour juvenile (30 days), young (1–2 years), adult (3.8–6years)and has been investigated under volitional swimming conditions. senility (7–8 years). During its life cycle, young S. hollandi’s Castro-Santos (2005) analysed the volitional swimming paired fins are orange and turn to grayish when fish grows up. behaviour of migratory fishes when traversing velocity ThespawningperiodforS. hollandi lasts from early May to barriers. Li et al. (2011b) studied the behaviour of the end of August. During this period, fish spawn in water Spinibarbus hollandi in a laboratory physical model. with slow velocity, and eggs are attached to gravel (Cai et al., The present paper has three objectives: 2007). S. hollandi does not have a migrating behaviour during its life cycle and spawns in its living area. 1. To improve the parameterization of the behaviour of the Most previous research on S. hollandi focussed on its food fish species Spinibarbus hollandi (Figure 1) and to habits and optimal conditions for spawning, while little is acquire data for validation of models for fish dynamics known at present about the fish’s preference with respect to by means of dedicated laboratory experiments under hydrological parameters. S. hollandi is an omnivorous fish volitional swimming conditions, with a broad feeding ability, including algae crustacean and 2. To develop and validate an individual-based model for aquatic insects. S. hollandi is also known to have a preference the dynamics of Spinibarbus hollandi, which is a further for running water with rocky substrate at the middle and refinement of the model reported by Li et al. (2010), bottom layers of rivers. Figure 1. Spinibarbus hollandi. Copyright © 2013 John Wiley & Sons, Ltd. Ecohydrol. 6, 586–597 (2013) 588 R. HAN et al. 3m 0.4m 16m Figure 2. D-ended recirculating flume with flat bottom and vertical baffle. Flow depth in the experiments was 1.0 m. Field observations on the Lijiang River have indicated a length of 43 m and a width that varies from 5 to 13 m. that the flow velocity and the substrate size are the dominant The flume was characterized by a natural heterogeneous hydrological parameters with respect to the behaviour and bathymetry installed in mortar. The flow conditions were dynamics of S. hollandi. Hence, the fish preference in created and controlled by a pump, a line of valves and a relation to these two parameters has been investigated in rectangular weir (Figure 3). two dedicated laboratory experiments under volitional In both flumes, velocities were measured with a SonTek swimming conditions. Acoustic Doppler Velocimeter at the centreline every 0.5 m in streamwise direction. Measurements were made in points situated at the bottom, at mid-depth and at the water surface in LABORATORY EXPERIMENTS the first flume, and at mid-depth and the water surface in the second flume.