The meso-scale habitat simulation model MesoHABSIM21 was used to assess the impact of low-head HPPs on fish populations. MesoHABSIM is a physical habitat modelling system developed for e-flow assessment and river channel restoration planning. It describes the utility of instream habitat conditions for aquatic fauna, allowing to simulate change in habitat quality and quantity in response to alterations of flow and river hydromorphology. Meso-scale habitats are defined as geomorphic units (GUs, such as pools, riflles, rapids, glides22) that can be used by species and life stages for a significant part of their diurnal routine23. A meso-habitat can be considered suitable or optimal when the configuration of hydraulic patterns, together with the attributes that provide shelter, create favourable conditions for survival and development of animals. MesoHABSIM approach is based on the aggregation of three models24:
- 1.
A hydromorphological model that describes the spatial mosaic of fish-relevant hydro-morphological features.
- 2.
A biological model describing the relationship between the presence and abundance of fish and the physical environment of the river.
- 3.
A habitat model quantifying the amounts, frequency and duration of the available habitat depending on the flow regime and local river morphology.
For the modelling, the time series of daily water discharge data in natural and altered (downstream HPPs) conditions were created for wet, normal and dry years in order to describe the habitat suitability in all possible hydrological conditions. Conditional habitat suitability criteria (CHSC) were developed to define the relationship between fish distribution and physical environment. Physical spatial measurements of river hydraulic and fish shelter attributes (current velocity, depth, discharge, sediments, woody debris, boulders, etc.) were conducted on a scale of mesohabitat during field surveys. SimStream plugin of QGIS25 was used to organize collected data for mesohabitat modelling.
Hydrological data and hydromorphological surveys
The daily time series of discharge data of three water gauging stations (WGSs; Bartuva-Skuodas, Venta-Leckava and Mūša-Ustukiai) were taken from the hydrological yearbook of the Lithuanian Hydrometeorological Service for the periods of 1970–2000 (period before construction of HPPs) and of 2001–2015 (period after construction). The WGSs are located downstream the selected HPPs, and their data were used for the assessment of the altered discharge conditions and the impact of HPPs on fish communities. Two additional WGSs of Minija River-Kartena (for the Bartuva and Venta rivers) and Nemunėlis River-Tabokinė (for the Mūša River) were chosen for the restoration of natural conditions of river discharge at case study sites according to the analogy method26. The selection of a river analogue was based on the same hydrological region, similar catchment area, similarity in physico-geographical and hydrometeorological characteristics, and absence of anthropogenic structures which interrupt the continuity of the river, e.g. dams. The regression equation between case study river and river-analogue was prepared using daily water discharge data of 1970–2000 (period before construction of HPPs). The natural regime of investigated rivers after construction of HPPs (2001–2015) was restored using regression equations. In this way, we obtain the annual hydrographs of the investigated rivers in natural and altered conditions. In order to evaluate the habitat suitability in all possible hydrological conditions, hydrographs were prepared for wet, normal and dry hydrological years (probability of 5, 50 and 95%, respectively), according to average discharge data in the period of 2001–2015.
Four different discharge values (from minimal to average) were defined for hydromorphological measurements in each site of the selected river. These discharges represented the minimum, average and maximum low flow discharges of 30 consecutive days (Q30_min, Q30_ave, Q30_max) in the warm period (May–September), and multi-annual mean water discharge (Qannual_mean) in 1970–2000 (before HPPs construction). According to the Lithuanian law, environmental flow (Qenv) is defined at each HPP as 80% or 95% probability of the mean minimum discharge of 30 consecutive days of the warm period11. A Laser Rangefinder (distance, inclination, azimuthal measurements) connected via Bluetooth with the field tablet was used for the mapping of hydromorphological units (HMUs, also called mesohabitats). The maps of HMUs polygons were digitized in the .shp format using MapStream plugin of QGIS25,27. The length of an analysed river reach was defined as 20 times the mean river width28. The depth and flow velocity measurements in each defined HMU were done using a propeller-type flow meter mounted on a wading rod. Depending on the polygon area, from 5 to 30 measurements were carried out in each HMU, while the measurement density (point/m2) was kept as constant as possible in each case study considering its size (on average one point per 6 m2 in the Bartuva, 20 m2 in the Mūša and 25 m2 in the Venta rivers).
The presence/absence of fish shelters and vegetation were assessed visually (see21 for details). All measurements were carried out as close as it is possible to four defined discharges (minimum low flow (Q30_min), average low flow (Q30_ave), maximum low flow (Q30_max) and annual mean (Qannual_mean)) of each selected case study (Table 1).
Fish data and conditional habitat suitability criteria
Four Cyprinidae fish species, which are common in cyprinid-dominated lowland rivers of Lithuania20, but differ in rheophily and reproduction habitat were selected for the assessment of HPPs impact: lithophilic rheophilic schneider Alburnoides bipunctatus and dace Leuciscus leuciscus, phyto-lithophilic eurytopic roach Rutilus rutilus, and diadromous lithophilic eurytopic vimba Vimba vimba (fish guilds according to29). Based on the classification of fish species in European rivers according to their overall resistance to habitat degradation30, the selected species also represent different guilds of tolerance capacity: schneider is intolerant species, dace and vimba are intermediate, and roach is tolerant31. These four species are all benthopelagic, and in this respect they are similar, but due to their different preferences for rheophilic conditions, spawning habitat and overall habitat quality, it was expected that their response to changes in flow conditions should also be different. Currently access for diadromous vimba to most rivers is limited by dams; therefore, habitat availability for vimba was modelled only in the Venta River, which is still accessible for this species and contains its spawning grounds.
To define conditional habitat suitability criteria (CHSC)21, the river monitoring database for 2008–2015 was used. Data on the physical, chemical and hydrological characteristics of river sites was collected by the Lithuanian Environmental Protection Agency (EPA). Fish monitoring and assessment of hydromorphological characteristics of the site at the time of sampling was carried out by the Nature Research Centre under agreement with EPA. Standardized single-pass electric fishing took place in mid-July–September on river sections with a minimum length of at least 10 times the wetted width (but not less than 50 m) using backpack pulse current electrofisher (type IG200-2; HANS GRASSL GmbH) with a maximum output of 800 V and a maximum power of 10.0 kW per pulse.
For CHSC construction, only river sites in natural conditions (from good to high ecological status according to the European Water Framework Directive) with a catchment area of 100–5000 km2 and sampled by wading were selected from the database. In total, 245 river sites were selected. 160 sites in 75 rivers (2/3 of the selected sites) were randomly selected and used to build CHSC. The remaining 85 locations in 53 rivers (1/3 of all locations) were used for calibration. Once the locations were selected, their depth and current velocity were classified into intervals of 0.15 m and 0.15 m s-1 following the MesoHABSIM protocol (up to 0.15, 0.15–0.3, 0.3–0.45, etc.). The preference of schneider, dace and roach for depth and current velocity was determined by their frequency of occurrence in each of the intervals. In order to minimize the impact of random catches, species were considered present only when the number of individuals exceeded 25th percentile of the number of individuals in all places where they were found. Species were considered abundant when the number of individuals was greater than the median abundance in all places where they were found. A species was considered present in a particular interval of depth or current velocity only when its frequency of occurrence was > 40%. Accordingly, a species was considered abundant only in those groups of depth and velocity where the number of individuals was greater than the median in more than 50% of the sites. The preference for the type of substrate and shelters was determined according to the analysis of these environmental variables in the river sites where the species should be present based on the criteria of depth and current velocity. According to the geomorphological and ecological definition of mesohabitat21,22, 10 m2 was considered the minimum surface that an HMU must have to be considered a suitable (species present) or optimal (species abundant) habitat for fish. When tested on an independent dataset (85 sites), CHSC were considered satisfactory for the presence of species when the species were present in > 60% of the sites meeting the criteria (total accuracy > 0.6). CHSC were considered satisfactory for the abundance of species when the species were present in > 60% of the sites meeting the abundance criteria and the abundance of individuals was higher than the median in at least 50% of these sites.
CHSC for vimba were selected by an expert judgement, analysing common features of the river sites where this species was observed. Migration of vimba to the majority of former spawning grounds is currently restricted by dams. Therefore, this species is constantly found in a limited number of rivers, in which vimba is present not only during spawning in spring, but is also common in specific habitats in summer and autumn.
For the validation of CHSC for schneider, dace and roach, a single-pass electric fishing was performed in 42 HMUs of 4 natural rivers (Minija, Dubysa, Šventoji and Merkys), in river stretches with a length of 150–400 m, a maximum depth up to 1.5 m, and a catchment size of 315–3040 km2, during the low flow season, with high transparency of water. Fish were sampled by wading by a team of 3 persons using a backpack pulse current unit of a similar type as for fish monitoring (IG200-2D; HANS GRASSL GmbH). CHSC verification for vimba was carried out only in 14 out of 42 HMUs, since this species is constantly found in only one of the natural rivers selected for verification. A single-pass electric fishing was also conducted in all HMUs which were identified in the studied river stretches below HPPs at the low flow. Fish sampling was accomplished by wading and using pulse current backpack electric fishing gear. A single-pass electric fishing strategy was used, as the CHSC criteria were also developed based on single-pass sampling data. Studies show that in most cases species composition and rank abundance of common species do not change significantly after the first pass32,33,34.
To assess the predictive performance of CHSC, correctly classified instances, sensitivity, specificity, and true skill statistic were calculated based on confusion matrix analysis35.
Assessment of HPPs impact
The habitat area available for the species was modelled at different discharges of rivers. The impact of HPPs on habitat availability was assessed based on the comparison of the modelled available habitat area (i) at reference conditions during a dry year, (ii) under HPPs functioning in dry, normal and wet years, and (iii) at environmental Qenv. The flow value that exceeded 97% of the time at reference conditions (Q97)36 during a dry year and the corresponding area of species habitat (expressed in m2, hereafter, the minimum threshold area) were used as common denominators. Deviation of temporal availability of suitable habitats for modelled fish species due to HPPs functioning at different flows was assessed based on relative increase in the cumulative continuous duration of days when the area of the habitat falls below the minimum threshold values (hereafter, the stress days alteration; SDA). SDA analysis is based on the assumption that minimum habitat availability is a limiting factor for fish species, and events occurring rarely in nature create stress to aquatic fauna and shape the community. Therefore, for the selected minimum habitat threshold (expressed in m2), the number of habitat stress days that occur under those conditions was calculated and used as a benchmark for comparative analysis using the SDA metric, (see e.g.28,36,37 for details). Finally, we normalize SDA values between 0 and 1 by using the index of temporal habitat availability (ITH) as it is described by Rinaldi et al.28.
The relative abundance of fish species that are common in the cyprinid-dominated rivers of Lithuania (the frequency of occurrence in the natural river sites is > 50%) was also compared in river reaches with natural (42 sites, 85 fishing occasions) and regulated (below HPPs; 20 sites, 39 fishing occasions) flows, which met at least good water quality criteria and fell within the same range of catchment size and slope as the rivers selected for modelling did. The sites were selected from the same river monitoring database for 2008–2015, which was used for selection of sites for CHSC development. The significance of identified differences was assessed using the Mann–Whitney U test.
Source: Ecology - nature.com
