First steps in the ecological flow determining for Latvian rivers

European rivers belong to ecosystems most impacted by hydromorphological alterations (EEA 2018), and without hydromorphological measures, it is not possible to achieve at least good ecological status which is the main purpose of the Water Framework Directive (WFD) 2000/60/EC (European Commission 2000). An unaltered hydrological regime is an important factor for the development of a healthy lotic environment (Sakaris 2013) and ecological flows must be determined in regulated rivers. According to the WFD CIS guidance document No. 31 (European Commission 2015), ecological flows are: ‘The amount of water required for the aquatic ecosystem to continue to thrive and provide the services we rely upon’. The relationship between biotic and abiotic factors in natural streams is still too poorly understood (Erba et al. 2006) and it also affects knowledge about the effects of disturbance in regulated streams. Although historically eutrophication has been considered a major pressure on European rivers, hydromorphological degradation today plays a major role (Baattrup-Pedersen et al. 2015). About 150 small hydropower plants (HPP) are operating on Latvian rivers and hydromorphological alterations are one of the most significant pressures within water bodies. According to the third-cycle Latvian River Basin Management Plans (LEGMC 2021), about 45% of river water bodies are affected by hydromorphological alterations and 30% of them are under significant hydromorphological pressure due to HPP. Of all biological quality elements recommended by WFD, fish is one of the most sensitive to hydromorphological alterations (Poff & Zimmerman 2010), including water level fluctuations (Schmutz et al. 2015). To achieve a better understanding of in-stream ecological processes, morphological indices, traditionally more widely used in river basin management plans, must be combined with indices sensitive to hydrological stressors caused by HPP (Schinegger et al. 2016). The scale of the study is challenging as it should be sufficient to assess not only the hydrological pressure but also the biota response where the catchment level may be insufficient. Studies have shown (Vezza et al. 2014; Wolter et al. 2016) that the mesohabitat scale is suitable to assess fish response to hydromorphological pressure and establish ecological flows. One of the approaches to link biology with hydrological alterations is to use meso-scale habitat models (Parasiewicz 2007) which simulate changes in habitat suitability for fish. Mesohabitat simulation system method MesoHABSIM is one of the most fully established mesohabitat simulation platforms (Wegscheider et al. 2020) which is most widely used in Central (Parasiewicz et al. 2018) and Southern Europe (Vezza et al. 2012), and the model has been tested also in the Baltic region (Virbickas et al. 2020; Akstinas et al. 2021).

Latvian legislation does not have a clear definition on what is an ecological flow and in what way it must be calculated. The definitions of ecological flow and minimum guaranteed flow in different legislation acts are sometimes contradictory. The concept of minimum guaranteed flow is more commonly used and it is mostly defined as a minimum or average of summer 30-day period low flow with 95% probability. This threshold is also mainly used to calculate ecological flow. Although ecological flow rate must be obligatory determined for hydroelectric power plants, in practice, in many cases, it is set equal to the minimum guaranteed flow rate. When a new permit is issued, ecological flow is recalculated, but there may be some cases where the ecological flow is missing or debatably calculated in the old permits. Hydropeaking is not regulated by Latvian legislation and was not addressed in this study. However, the water level monitoring results show its frequency and amplitude that are critical in terms of ecosystem and river banks erosion process.

The first efforts of the application of MesoHABSIM in Latvia started in 2017 and by now it has been applied to seven rivers.

The aim of this study is to analyse the results of MesoHABSIM modelling in these rivers, estimate the recommended ecological flows and compare them with the ecological flow currently stated in water use permits.

Water flow data series for the period 1961–2018 were used for the assessment of hydrological regime regulation by HPP during different seasons as well as for the river habitat modelling. Two daily flow data series (in reference and altered conditions) are required in the modelling procedures. Water flow data series before HPP construction is used to calculate flow in the reference conditions. The station-analogue flow data series were used for the ungauged rivers. Flow data series of rivers located in similar hydromorphological and climatological conditions (terrain, river feeding, watershed area, amount of precipitation, etc.) were taken as the analog flow. In this case, flow data of ungauged river are modelled on the base of analog river’ data, using the ratio of basin areas of pilot rivers and river-analog. For calculation of altered flow, information about HPP daily hydropower production was used. This calculation is based on the power equation: H*Q*9.81*k, where W – power, kW; H – Head, m; Flow – l/s, 9.81 m/s² gravity constant; K – combined turbine and generator efficiency (approximately 0.6). The studied river site and HPP characteristics are given in Table 1.

Two rheophilic fish species potentially vulnerable to alterations of hydrological regime were selected for modelling of ecological flow – sea trout/brown trout Salmo trutta for salmonid waters and chub Squalius cephalus for cyprinid waters. Trout is abundant species in small- and middle-sized salmonid streams while chub – in middle-sized and large cyprinid rivers (Birzaks 2013). Both species have also been captured in all salmonid or cyprinid rivers analysed in this paper (unpublished data of institute ‘BIOR’).

Meso-scale habitat simulation model MesoHABSIM was used for habitat suitability calculations at different flows. It is based on habitat availability for selected fish species during different hydrological conditions (Vezza et al. 2012).

MesoHABSIM consists of three separate sub-models:

1. Fish conditional model: fish habitat model which describes relationships between the abundance of selected fish species and the abiotic environment of the river (depth, stream velocity, substrate composition, presence of boulders, woody debris or in-stream vegetation, etc.).

2. Hydrological data: flow time series in reference (natural) and altered (impacted by HPP) conditions.

3. Hydromorphic unit (HMU) data: HMU as polygons and hydromorphological data as points based on field measurements, including river depth, channel substrate and stream velocity.

Sim-Stream computer model tool in QGIS (Zanin et al. 2016) was used to implement MesoHABSIM model. Sim-Stream software combines all three parts of MesoHABSIM (fish model, hydrological data series and HMU) and simulates physical habitat suitability at different flow conditions.

For more details about MesoHABSIM model application, please see other articles (Parasiewicz 2007; Parasiewicz et al. 2018).

Fish conditional models – used in this study – were developed in Interreg V-A Latvia–Lithuania Programme 2014–2020 project ‘Ecological flow estimation in Latvian – Lithuanian Transboundary river basins (ECOFLOW)’, LLI-249.

Habitat flow-rating curves and habitat suitability maps were done using SimStream software tool in QuantumGIS. Fish sampling was done at the studied river reaches but no direct habitat suitability validation measures were done within this study.

Hydromorphological unit mapping and field works were done in 2017 (Venta RBD), 2018 (Lielupe RBD) and 2019 (Vaidava River). Surveys were done between March and October, no surveys were done in the winter season due to the ice. Hydromorphological type-specific, meso-scale river stretches were selected downstream of each studied HPP. Depending on river size, these river stretches were 100–500 m long. Only natural sites without channelisation were selected in order to assess the ecological impact of water level alterations below HPP. Each river stretch was divided into hydromorphological units (HMU), which were mapped at multiple flow conditions. HMU can be described as lotic mesohabitats (riffles, rapids, glides, pools). HMU was mapped as polygons, which allows assessing changes in habitat areas under different water levels. Flow velocity, water depth and channel substrate were measured at least in seven points within each HMU. Depending on river size and the number of HMU units, 40–100 measurements were done at each site. Field cartography was done using a field computer, a TruPulse 360 range finder and ESRI ArcPad software. Rivers were surveyed four times during different flow conditions: summer low flow minimum (Q_{30_min}), low flow average (Q_{30_avg}), low flow maximum (Q_{30_max}) and annual average flow (Q_{annual_avg}).

Different habitat–flow-rating curves were observed for cyprinid fish rivers. Optimal flow was achieved at lower flows, corresponding to Q_{30_avg} (Figure 7). We did not observe any habitat decrease trends with increasing water discharge and flow velocity, because chub is relatively tolerant to an increase of stream velocity and water depth is the most limiting factor. Auce River was relatively unsuitable for chub because modelled habitat suitability requires that water depth is at least 30 cm for juveniles and 60 cm for adults. Because of HPP operation and water storage, it was only possible to assess this river in mostly low flow conditions.

This research is the first attempt in Latvia to set biologically significant type-specific ecological flow values using fish as bioindicators. Our results show that different fish species have different modelled responses to hydromorphological alterations due to operating HPP. In accordance with typical type-specific river fish communities found in Latvia (Birzaks 2013), we chose brown trout and chub as examples for determining ecological flow. In the future, other fish species than chub must be used because this large-body fish may not be found in upstream sites of rivers. Because of its large size, chub is also very sensitive to water shortage and below HPP it may be replaced with smaller-body fish (Kubecka et al. 1997). As shown in Figures 7 and 8, habitat availability differs significantly in salmonid and cyprinid rivers indicating that ecological flows in Latvia must be calculated separately for different regions and river types. We found regional differences in habitat suitability trends in salmonid rivers which are in line with hydrological districts in Latvia. In comparison to examples from Venta RBD, our calculations show that Vaidava River has a larger bed slope and discharge, which is too large and unsuitable for fish communities when reaching a certain threshold. Due to the limited number of studied sites, it was not possible to analyse possible regional differences for cyprinid river types. The optimum flow obtained for cyprinid rivers is very close to the average low flow which can be used as a threshold to define ecological flows for cyprinid rivers. A similar approach was proposed also in Lithuania (Virbickas et al. 2020). For salmonid rivers, the optimum flow was higher and closer to low flow maximum flow which can be used as a threshold for salmonid rivers. On average, the ecological discharge must be raised more than three times (Table 2), leading to significant energy losses but no estimations are available about biodiversity losses due to a too low water flow below HPP. According to the third-cycle Latvian River Basin Management Plans (LEGMC 2021), the implementation of proposed ecological flows may cost about 1% of the long-term profit of HPP.

In general, ecological flows calculated in our study significantly differ from ecological or guaranteed flows specified in HPP water use permits (Table 2). The main difference between the two thresholds is that in water use permits ecological flow values were determined using only hydrological calculations, but in this study ecological flow values were set using fish's ecological response to changing water levels. All studied HPP have ecological flows specified in water use permits, but for three of them, ecological flow is equal to guaranteed flow which is very low water discharge. Only three HPP (Karva, Grube and Rundale HPP) have an existing ecological flow which partly is in line with the ecological flow calculated within this study. The ecological flow of the Vanka River is extremely low and, according to our results, it should be increased at least five times to provide suitable habitat for fish. Although current legislation theoretically regulates the calculation of ecological flow, in practice the calculation approaches are different. The ecological flow rate is most often determined by fisheries expertise and has been calculated as half of the average flow rate during the two months of summer low flow (Abersons & Jekabsone 2018). The results of this study show that in most cases the currently used ecological flow approach does not guarantee the achievement of the objectives of WFD.

Our proposal is to calculate the ecological flow regime, not to set only one ecological flow value. This approach is proposed also in Spain where optimum flow is calculated separately for juveniles and adults (Jalon 2003). Using a flow regime instead of a single ecological flow value has several advantages. The flow regime takes into account fish bioperiods: adult spawning in winter (from November to May) and juvenile growing in summer (from June to October). The flow regime may provide an ecologically friendly environment not only for fish but also for other aquatic organisms like freshwater mussels (Sousa et al. 2020). The results of this study are also in line with other studies (Ceola et al. 2018; Parasiewicz et al. 2018), indicating that ecological flows possibly can be calculated at the regional level using the grouping of rivers with comparable hydromorphological features.

Lielupe river basin district is the most hydromorphologically impacted region in Latvia (LEGMC 2021) and rivers usually have multiple pressures (channelisation, eutrophication, dams) affecting them which may cause a problem to determine where hydrological pressures caused by operating HPP are the most important. Additional site-specific studies must be done in these heavily impacted and modified rivers (Parasiewicz et al. 2018) to validate our findings and set necessary thresholds to achieve at least good ecological status or potential. Two of our cyprinid case studies (rivers Islice and Auce) are heavily impacted by eutrophication which may affect fish assemblages (Sutela et al. 2010) and results of ecological response to water level alterations.

The currently existing ecological flows, provided in water use permits, do not completely support the sustainability of aquatic ecosystems for most of the studied rivers. Although Latvia is a small country in terms of area, regional differences can be observed in the hydrological regime of rivers. That potentially affects the calculation of the ecological flow. For salmonid and cyprinid rivers, the ecological flow regime must be calculated differently because salmonid rivers are more vulnerable to artificial low flows. Different fish live stages have different requirements for flow intensity and therefore ecological flow regime must be calculated as a range, taking into account flow preferences for juveniles and adults. Additional research must be conducted in all four river basin districts, taking into account salmonid and cyprinid fish waters.

We are grateful to our colleagues of Interreg V-A Latvia–Lithuania Programme 2014–2020 project ‘Ecological flow estimation in Latvian – Lithuanian Transboundary river basins (ECOFLOW)’, LLI-249 Janis Birzaks, Tomas Virbickas and Paolo Vezza for the introduction of MesoHABSIM and great effort in the development of fish conditional models used in this paper.

JJ, KA, TK conceived of the presented idea; JJ, KA, TK, MT helped supervise the field works; JJ, KA, TK analysed the data; JJ, KA wrote the original draft; JJ, KA, TK, MT reviewed and edited the article. All authors have read and agreed to the published version of the manuscript.

Data cannot be made publicly available; readers should contact the corresponding author for details.

The authors declare there is no conflict.