The Romanian Ecological Flow Method, RoEflow, Developed in Line with the EU Water Framework Directive. Concept and Case Studies

Abstract

The overall purpose of the research is to develop a method to compute ecological flows in line with the EU Water Framework Directive (Directive 2000/60/EC) for the whole Romanian territory, for a variety of hydrological, morphological and ecological conditions. The method has three components: a Quantity component, a Dynamic component, and a Real-time operation component. The Quantity component is a hydrological method with elements of the aquatic fauna habitat indirectly linked to biological organisms based on the current Romanian knowledge on the linkages between hydrology and aquatic biology. The Dynamic and Real-time operation components are related to the hydrological forecast. The method is practical, robust and easy to apply. The concept and the ideas use the hydrological forecast to ensure the water dynamics required by the Water Framework Directive, and to develop the quantitative component, keeping in mind that putting it into practice might have importance for a broader audience. In order to better highlight the concept, the paper shows three practical examples of the RoEflow method’s application.

1. Introduction

For many rivers, the hydrological regimes are being modified due to dams, weirs, abstractions for water supply, industry and agriculture, hydropower and different structures for flood control. Besides the diminishment of the flows, there is an impact on the seasonality, magnitude of flow and flood frequency, with the effect of reducing or eliminating the link between the river channel and its floodplain, and on river ecosystem services [1]. In addition, there are global land use and climate change as additional stress factors for rivers [2].

The water demands continue to grow in many countries, including those situated in semi-arid regions, which are already affected by medium or high-water stress [3]. Even in these conditions, the hydrological regimes must not be significantly altered in terms of quantity and dynamics. In this sense, the ecological flow was defined for the aim “to protect or restore the integrity and health of river ecosystems (functions and processes)” [4,5]. The recommendations for the computation of ecological flows have been grounded in the relationships among the flow variables and ecological responses to natural flow variability and flow alterations from the natural or historic baseline [6,7].

The European Guidance Document No. 31—Ecological flows, in the implementation of the Water Framework Directive [8], defines ecological flows as a hydrological regime that ensures the achievement of the environmental objectives of the Water Framework Directive (WFD) [9] (the non-deterioration of ecological status, good ecological status, the objectives of protected areas where relevant) in natural surface water bodies.

In addition, the Brisbane Declaration and Global Action Agenda on Environmental Flows (2018) provides recommendations to support and promote ecological flow implementation as a central element of sustainable water resource management [10].

Numerous methods for the assessment of ecological flows have been developed over time: hydrological, hydraulic-habitat and holistic methodologies [11]. Among these, the most used methods in the world are hydrologically based, having as advantages the availability of the local time series of hydrological data (measured or estimated), thus saving time and money for data collection [8,11].

Hydrological methods are based on the assessment of the natural flow regime, which involves ensuring the processes and conditions that will maintain the native habitats and species [12].

The disruption of a natural flow regime (flows that are too low or high at certain times) can have a negative impact on the structures and functioning of aquatic ecosystems and ecosystem services.

The most comprehensive methodologies assume that the full range of natural variability in the hydrological regime is necessary to conserve aquatic ecosystems. The characterization of a natural range of variability, on the short-term features of the flow regime and on the natural flow variability over longer periods, aims to define e-flows which are able to maintain an adequate habitat for aquatic biota [6].

Many hydrological methods for the calculation of ecological flow consider the natural regime, and in addition, the requirements for fish habitats, based on hydraulic calculations, including the breeding habitat and periods (e.g., methods used in Austria, Italy, Slovenia, China, etc.). For example, in China, a case study recommends two periods for computing ecological flows, including the spawning period (from April to September) and the normal period (from October to March of the following year), considering both the runoff and fish habitats [13].

As such, many ways to compute ecological flows have been derived over time. However, a compromise is usually reached between human values and environmental benefits considering the socioeconomic impacts when the ecological flow is implemented [14].

In many cases, environmental flow implementation is insufficient and water abstraction effects are underestimated [15].

The current paper shows a method for the computation of ecological flow. The old Romanian methods are exclusively hydrological methods with a single constant value all over the year. The Romanian ecological flow (RoEflow) method shown within this paper is a hydrological method with elements of aquatic fauna habitat, ensuring inter- and intra-annual variability. The RoEflow method is based on the current Romanian knowledge on the linkages between the hydrological regime and the aquatic habitat.

The dynamic and real-time components are original. The current hydrological monitoring network and forecasting system support this method. It is robust and easy to apply for different hydrological conditions and different morphological features (mountains, hills, lowland), and implicitly different ecosystems. The RoEflow method was recently approved by a governmental decision. The stakeholders’ consultation was carried out during the development of the method. The current paper describes the concept of the RoEflow method and shows three case studies.

2. Materials and Methods

The European Guidance Document No. 31—Ecological flows, in the implementation of the Water Framework Directive, recommends that member states develop “a conceptual definition of ecological flows with a clear reference to both flow quantity and dynamics and to their consistency with the environmental objectives required under the Water Framework Directive”. The same guidance document specifies that the values for the ecological flows should be derived using natural flows. The RoEflow method (Figure 1) takes into consideration these recommendations.

The RoEflow method has three components: (A) a Quantity component; (B) a Dynamic component, and (C) a Real-time operation component. Component A was developed in a research study within the National Institute of Hydrology and Water Management, Bucharest, Romania, and components B and C are the original contributions of this paper.

A. Quantity component

This component describes the steps to derive the quantity of water consistent with the environmental objectives required under Directive 2000/60/EC, named the Water Framework Directive.

The hydrological parameters used within the A component are explained below: $Q_{i,j}^{month. a}$ is the monthly average discharge, i.e., the volume of water that flows, on average, through a given cross-section of a river during month (j) of the year (i).

$Q_j^{m month. a}$ is the multiannual monthly average discharge, i.e., the monthly water volume that flows, on average, over many years, through a given cross-section of a river.

The multiannual monthly average discharge for month j is

$$Q_j^{m month. a}=\frac{{{\displaystyle \sum }}_{i=1}^n{Q}_{i,j}^{month. a}}{n}, n=30 \mathrm{years}$$

(1)

$Q^{m a}$ is the multiannual average discharge, i.e., the water volume that flows, on average, over many years, through a given cross-section of a river.

$$Q^{m a}=\frac{{{\displaystyle \sum }}_{i=1}^n{Q}_{i }^{a a}}{n}$$

(2)

where: $Q_i^{a a}$ is the average annual discharge in year I, and is computed with the following equation:

$$Q_i^{a a}=\frac{{{\displaystyle \sum }}_{j=1}^{12}{Q}_{i,j}^{month.a}}{12}$$

(3)

The above discharges are computed using natural or naturalized/estimated flows, for a 30-year period (recent records), in order to capture dry, rainy and normal hydrological years, and the influence of climate change. A series of 30 years is in line with the current hydrological practice recommended by the World Meteorological Organization.

The hydrological parameters were chosen as monthly average discharges due to the fact that they are easy to derive over a large territory, and easy to naturalize. At the same time, there is a unitary approach at the level of a country. The hydrological parameters of the method (monthly average discharges) can be derived in controlled river basins, as well as in uncontrolled catchments with scarce (short data series) or lacking measurement records using indirect methods (e.g., regionalization methods, modelling). The use of hydrological regionalization methods is a common hydrological practice in Romania.

A.1. Hydrological optimum.

The Water Framework Directive requires, implicitly, that we identify links between the flow regime and the biological quality elements (fish, benthic invertebrates, phytobenthos, aquatic macrophytes or phytoplankton). The identification of the tolerance limits of the aquatic organisms against certain parameters of the aquatic environment (temperature, chemicals, etc.) means that the identification of the ranges for the favorable (or ecological optimum) domain is required.

The “ecological optimum” range corresponds to the domain in which species live on a reduced consumption of energy and matter [12]. It is the domain in which the habitat conditions for the reproduction of the species are ensured. The ecological optimum has several components: climatic, chemical and physical components, and the hydrological one, etc. The “hydrological optimum” is the domain that provides the necessary conditions for the reproduction and development of the aquatic organisms, i.e., the ecological flow.

Among the aquatic organisms, the fish were chosen in order to find the hydrological optimum regime (ecological flow), keeping in mind the following points:

-

The fish fauna is located on the top of the trophic chains within the aquatic ecosystem; if there are fish in a lotic aquatic ecosystem, there are other biological elements (benthic invertebrates, phytobenthos, aquatic macrophytes or phytoplankton) which are the sources of food or habitat for the fish.

-

The fish fauna is considered a good indicator of the global anthropogenic impact.

-

The fish communities consist of several trophodynamic modules (omnivores, phytophagous, predators) and integrate the response of other components of the ecosystem through their position in the trophic chains and longer life span compared to invertebrates.

It should be considered that the habitat needs of the various fish species are different, depending on the evolutionary stage of the life cycle and species features: wintering, feeding and reproduction.

The assumptions regarding the Romanian Ecological Flow computation are mentioned below:

-

The fish need a complete range of the natural variability of the hydrological regime.

-

The natural flow is the support for the habitat requirements for the dominant fish species (for each river typology) that have been preserved over time (they existed before 1964, i.e., they were mentioned in the Treaty regarding the fauna of Romania, Volume 13—Fishes, published by academician PhD. Petre Mihai Bănărescu [16]), and still exist now according to the results obtained in the monitoring campaigns carried out by the National Administration “Romanian Waters” authority.

-

The need for a larger amount of water during the fish breeding periods and in protected areas.

The “hydrological optimum” described theoretically above is defined as the domain ranging between $mi{n}_j\left\{mi{n}_i\left\{Q_{i,j}^{month. a}\right\}\right\}$ and $ma{x}_j\left\{mi{n}_i\left\{Q_{i,j}^{month. a}\right\}\right\},$ where: i is the month, j is the year, and $Q_{i,j}^{month. a}$ is the monthly average discharge in year i and month j, computed using a long series of natural flows (1950–2013). In general, the ecological flows should vary between these limits (peaks above the upper limit are allowed). Some issues related to the validation of the hydrological optimum domain are shown within Appendix A.

A.2. Equations and coefficients linked with elements of aquatic fauna habitats.

The RoEFlow method uses the monthly average discharge in the natural regime for 30 years as the standard period. The natural flow regime is recommended by the European Guidance Document No. 31—Ecological flows in the implementation of the Water Framework Directive (p. 4): ”Therefore consideration of ecological flows should be included in national frameworks, including binding ones as appropriate, referring clearly to the different components of the natural flow regime (and not only to minimum flow) […]”. Flow variability was ensured by defining 12 different values that involve the natural flow variability with low, average and high flows. The ecological flows have the same “shape” as the natural hydrograph. The RoEFlow method links the hydrological regime with elements of aquatic fauna habitat through the coefficients and their calibration. Details on the method calibration are shown in the Appendix A.

The equations for the monthly ecological flows in a certain river cross-section are shown below:

$$Q_{eco,j}={\beta }_1·Q_j^{m month.a}, \mathrm{when} Q_j^{m month.a}\le Q_{m a}$$

(4)

$$Q_{eco,j}={\beta }_2·Q_j^{m month.a}, \mathrm{when} Q_j^{m month.a}>Q_{m a}$$

(5)

${\beta }_1=0.25-0.35$ for mountain and hill river typological groups; ${\beta }_1=0.20-0.30$ for lowland river typological groups.

${\beta }_2=0.25-0.35$ for mountain and hill river typological groups; ${\beta }_2=0.25-0.30$ for lowland river typological groups.

The coefficients ${\beta }_1$ and ${\beta }_2$ depend on the river typological groups (see

Table A1. The characteristics of the river typological groups of Romanian typologies [20].

River Typological Groups	Altitudes (m)	Mean Rainfall (mm/year)	Mean Air Temperature (°C)	Lithological Structure	Potential Fish Fauna *
Mountain	>500	600–1400	−2–+9	blocks, boulders, gravel, sand	trout, grayling, chub, nase
Hill	200–500	500–700	8–10	gravel, sand	chub, barbell, nase
Lowland	<200	400–600	9–11 (>11)	sand, muddy, muddy clay, clay	carp, nase chub, perch, barbell

* No fish for the intermittent/temporary river typology.

of Appendix A). They consider the dominant fish species, ecosystem characteristics, biological activities (reproduction) and the period in which they occur. If the analysed cross-section is in a protected area, the maximum values of these coefficients will be used, taking into account that the protected areas require a higher level of protection for aquatic organisms, as well as for species for conservative interest. Some issues related to the calibration and validation of the ${\beta }_1$ and ${\beta }_2$ coefficients are highlighted within Appendix A.

The ecological flows are computed for each month of the year using Equations (4) and (5). The 12 values obtained, $\mathrm{i}.\mathrm{e}., Q_{eco,j}$, are grouped into three types of hydrological regime (low, average, high flow) as follows:

$$\mathbf{low} \mathbf{flow}: Q_{eco low flow }=max \left(Q_{95\%}; min\left(Q_{eco,j}\right)\right)$$

(6)

$$\mathbf{average} \mathbf{flow}: Q_{eco average flow }=median\left(Q_{eco,j}\right)$$

(7)

$$\mathbf{high} \mathbf{flow}: Q_{eco high flow }=average of the highest 4 values of Q_{eco,j}$$

(8)

where $Q_{95\%}$ represents the yearly minimum monthly mean discharge with a 95% probability of occurrence.

The $Q_{95\% }$ value was chosen as the minimum threshold value below which the hydrological regime characteristics for small waters cannot be reduced without a significant influence on the connection between the river and the groundwater aquifer (the connection to groundwater bodies is required in Annex V of WFD). Furthermore, in Romania, $Q_{95\% }$ was considered the dilution flow for water quality protection (e.g., in case accidental water pollution occurred), and it has been used for a long time in water management in Romania.

B. Dynamic component assessment

The RoEflow method has a dynamic component related to hydrological forecasting, which is one of the original contributions of this paper. The concept of component B was developed to be linked with the current practice of hydrological forecasting in Romania. The dynamic component of the method is described below.

B1. Matrix with the monthly forecasted discharge, time, and class of forecast

The method has five classes of forecast related to the average multiannual monthly flow $Q_j^{m month.a}$, defined as follows (Equations (9)–(13)):

$$\mathrm{Class} \mathrm{of} \mathrm{forecast} “>100\%”—Q_{forecasted}^{month.}=Q_j^{m month.a}$$

(9)

$$\mathrm{Class} \mathrm{of} \mathrm{forecast} “80–100\%”—Q_{forecasted}^{month.}=0.8 ·Q_j^{m month.a}$$

(10)

$$\mathrm{Class} \mathrm{of} \mathrm{forecast} “50–80\%”—Q_{forecasted}^{month.}=0.50 ·Q_j^{m month.a}$$

(11)

$$\mathrm{Class} \mathrm{of} \mathrm{forecast} “30–50\%”—Q_{forecasted}^{month.}=0.30 ·Q_j^{m month.a}$$

(12)

$$\mathrm{Class} \mathrm{of} \mathrm{forecast} “<30\%” Q_{forecasted}^{month.}=0.15 ·Q_j^{m month.a}$$

(13)

B2. Matrix with the minimum monthly forecasted discharge, i.e., $Q_{forecasted}^{min month.}$, time and class of the forecast.

The ecological flows must be assured at all time; therefore, they should be linked with the minimum monthly forecasted flows $Q_{forecasted}^{min month.}$ (Equation (14)).

$$Q_{forecasted}^{min month.}=c· Q_{forecasted}^{month.}, \mathrm{where} c=0.35–0.5$$

(14)

The “c” coefficient results from the analysis of the ratio between the monthly minimum discharge (instantaneous values) and the monthly average discharge ($Q_{i,j}^{month. a}$) for a series of representative gauging stations with a natural flow regime.

B3. Matrix with $Q_{eco high flow}$, $Q_{eco average flow}$, $Q_{eco low flow}, Q_{forecasted}^{min month.}$, time and the class of the forecast.

This step is linked with the results of A, i.e., the Quantity assessment (see Figure 1).

The values of $Q_{forecasted}^{min month.}$ from each class of forecast are compared to the ecological flow set for each hydrological regime type ($Q_{eco high flow}$, $Q_{eco average flow}$, $Q_{eco low flow}$). The appropriate values corresponding to low, average and high flows are selected as follows (Equations (15)–(17)):

$$\mathrm{High} \mathrm{flow}: Q_{forecasted}^{min month.} \ge Q_{eco high flow}$$

(15)

$$\mathrm{Average} \mathrm{flow}: Q_{eco average flow}\le Q_{forecasted}^{min month.}<Q_{eco high flow}$$

(16)

$$\mathrm{Low} \mathrm{flow}: Q_{forecasted}^{min month.}<Q_{eco average flow}$$

(17)

B4. Reduce the quantity from $Q_{eco high flow}$ to $Q_{eco average flow}$, or from $Q_{eco average flow}$ to $Q_{eco low flow}$, outside the fish breading period.

The breeding period (the specific months for reproduction) of the fish species related to the river typological groups is identified. The details related to the river typological groups are given in Table A1 of Appendix A.

Due to the fact that the long-term forecasts are associated with a high degree of uncertainty, for the months outside the breeding period, the ecological flow may be reduced from $Q_{eco high flow}$ to $Q_{eco average flow}$, or from $Q_{eco average flow}$ to $Q_{eco low flow}$(a deviation is taken into account).

C. Real-time operation component.

The RoEflow method has a real-time operation component, which is one of the original contributions of this paper. This component assures the flow dynamics, month by month, of the river downstream of the analyzed cross-section.

In practice, according to the monthly hydrological forecasting, which is available on the National Institute of Hydrology and Water Management (NIHWM), Bucharest, Romania website [17], the ecological flow values corresponding to the hydrological regime ($Q_{eco high flow}$, $Q_{eco average flow}$, or $Q_{eco low flow}$) related to the forecasted hydrological regime for the area/cross-section of concern will be used.

The above-mentioned method is solely for the computation of ecological flows, and is used for natural water bodies as well as for heavily modified water bodies, i.e., HMWBs (for the definition, see article 4(3) of the Water Framework Directive, WFD). The implementation into practice is another important issue. It should be carried out considering the particularities of HMWBs. The exact values for the delivered ecological flow downstream dams should be established with respect to the provisions within the WFD related to HMWBs.

3. Case Studies

The above-described method was applied to three cross-sections in Romania (Figure 2). The cross-sections are spread within Romania. Their hydrological, morphological and ecological conditions are diverse. The first case study is detailed below.

The first cross-section is located on the Valea Cormaia river within the Someș river basin, downstream of a small water intake within a natural water body. The water intake is used for a fish farm. The watershed of the cross-section has an area of 70 km², a mean altitude of 1246 m, a mean annual temperature of about 7 °C (1961–2013), and a mean annual rainfall of 1020 mm/year (1961–2013). There are no sources of pollution, and the land use is predominantly natural: 44.8% forest, 55% semi-natural areas (scrub and/or herbaceous vegetation associations) and 0.2% open spaces with little or no vegetation.

There are protected sites within the water body [18]. The river cross-section belongs to the RO01 river typology, in the mountain typological group (see Table A1 of Appendix A). According to Romanian river typologies, the fish species are trout, grayling and chub (Table A1 of Appendix A). Their breeding periods are March–June and October–December [16].

The second cross-section is located on the Teleajen River, within the Prahova River basin, downstream of a hydropower plant water intake, within a natural water body. There are no protected sites within the water body [18]. The watershed of the cross-section has an area of 769 km², a mean altitude of 784 m and a mean annual rainfall 822 mm/year (1961–2013). The river cross-section belongs to the RO05 river typology, in the hill typological group (see Table A1 of Appendix A). According to Romanian river typologies, the fish species are nase and barbell (Table A1 of Appendix A). Their breeding periods are April–June [16].

The third cross-section is located on the the Baboia River, within the Desnaţui River basin, downstream of an irrigation water intake, within a natural water body. There are protected sites within the water body [18]. The watershed of the cross-section has an area of 369 km², a mean altitude of 162 m and a mean annual rainfall 577 mm/year (1961–2013). The river cross-section belongs to the RO06 river typology, in the lowland typological group (see Table A1 of Appendix A). According to Romanian river typologies, the fish species are chub, perch and carp (Table A1 of Appendix A). Their breeding periods are March–June [16].

The input hydrological data consist in the monthly average discharge series in the natural regime (as recommended by the European guidance document) for a recent period of 30 years (1986–2015) in order to cover the whole spectrum of the flow dynamics. Thirty years is the standard number of years according to the World Meteorological Organization (used also for the meteorological series), which captures the recent trend of climatic changes and real potential data. The data set was provided by the National Institute of Hydrology and Water Management, Bucharest, Romania.

4. Results

4.1. Results of A—Quantity Component

The presentation of the results follows steps A2, B1, B2, B3 and B4, as described in Section 2 and shown in Figure 1.

The first cross-section analyzed is located within the mountain typological group. As shown at point A2, Section 2, for the mountain typological group, the coefficients ${\beta }_1$ and ${\beta }_2$ vary in the range 0.25–0.35 (see the explanation for Equations (4) and (5)). Because the water body is located in a protected area, the maximum values of these coefficients were selected (${\beta }_{1 }$=${\beta }_{2 }$= 0.35), aiming at a higher protection for aquatic environment. Using the Equations (4) and (5), the derived monthly distribution of the ecological flows is shown in Figure 3.

The results using Equations (6)–(8), namely the ecological flow for high (blue color), average (green color), and low (yellow color) water, are presented in the

Table 1. The input data and the results of A—Quantity component.

Month/Q (m3/s)	Jan.	Feb.	Mar.	Apr.	May	Jun.	Jul.	Aug.	Sep.	Oct.	Nov.	Dec.
$Q_{eco,j}$	0.361	0.341	0.829	1.486	1.070	0.595	0.461	0.281	0.371	0.416	0.505	0.495
$Q_j^{m month. a}$	1.03	0.97	2.37	4.25	3.06	1.70	1.32	0.80	1.06	1.19	1.44	1.41
$Q_{m a }$	1.72
$Q_{95\% }$	0.245
$Q_{eco low flow }$	0.281
$Q_{eco average flow }$	0.478
$Q_{eco high flow }$	0.995

. It contains the multiannual monthly average discharge ($Q_j^{m month. a}$) and the multiannual average discharge ($Q_{m a }$) as input data for the computation of the monthly ecological flows ($Q_{eco,j}$) using Equations (4) and (5). The $Q_{95\% }$ value input in Equation (6) is also mentioned within the table.

The procedure for the Quantity component was followed for the second and third case study. The results are highlighted in Figure 4 and Figure 5.

4.2. Results of B—Dynamic Component

The Dynamic component has a sequence of four steps (see Section 2, B. Dynamic component assessment and Figure 1). Therefore, Equations (9)–(17) were applied. For the first case study, the results of steps B1, B2 and B3 are shown in

Table 2. Monthly forecasted flows—$Q_{forecasted}^{month.}$, depending on the class of forecast (m³/s).

Class of Forecast	Jan.	Feb.	Mar.	Apr.	May	Jun.	Jul.	Aug.	Sep.	Oct.	Nov.	Dec.
>100%	1.031	0.974	2.368	4.246	3.058	1.701	1.318	0.804	1.061	1.190	1.443	1.414
80–100%	0.825	0.779	1.895	3.397	2.447	1.361	1.055	0.643	0.848	0.952	1.154	1.131
50–80%	0.516	0.487	1.184	2.123	1.529	0.851	0.659	0.402	0.530	0.595	0.721	0.707
30–50%	0.309	0.292	0.711	1.274	0.917	0.510	0.396	0.241	0.318	0.357	0.433	0.424
<30%	0.155	0.146	0.355	0.637	0.459	0.255	0.198	0.121	0.159	0.178	0.216	0.212

,

Table 3. Minimum monthly forecasted flows, $Q_{forecasted}^{min month.}$, depending on the class of forecast (m³/s).

Class of Forecast	Jan.	Feb.	Mar.	Apr.	May	Jun.	Jul.	Aug.	Sep.	Oct.	Nov.	Dec.
>100%	0.361	0.341	0.829	1.486	1.070	0.595	0.461	0.281	0.371	0.416	0.505	0.495
80–100%	0.330	0.312	0.758	1.359	0.979	0.544	0.422	0.257	0.339	0.381	0.462	0.452
50–80%	0.258	0.243	0.592	1.062	0.765	0.425	0.330	0.201	0.265	0.297	0.361	0.353
30–50%	0.155	0.146	0.355	0.637	0.459	0.255	0.198	0.121	0.159	0.178	0.216	0.212
<30%	0.077	0.073	0.178	0.318	0.229	0.128	0.099	0.060	0.080	0.089	0.108	0.106

and

Table 4. $Q_{eco high flow}$, $Q_{eco average flow}$ and $Q_{eco low flow}$, depending on the class of forecast (m³/s).

Class of Forecast	Jan.	Feb.	Mar.	Apr.	May	Jun.	Jul.	Aug.	Sep.	Oct.	Nov.	Dec.
>100%	0.281	0.281	0.478	0.995	0.995	0.478	0.281	0.281	0.281	0.281	0.478	0.478
80–100%	0.281	0.281	0.478	0.995	0.478	0.478	0.281	0.281	0.281	0.281	0.281	0.281
50–80%	0.281	0.281	0.478	0.995	0.478	0.281	0.281	0.281	0.281	0.281	0.281	0.281
30–50%	0.281	0.281	0.281	0.478	0.281	0.281	0.281	0.281	0.281	0.281	0.281	0.281
<30%	0.281	0.281	0.281	0.281	0.281	0.281	0.281	0.281	0.281	0.281	0.281	0.281

$Q_{eco high flow}$ is marked in blue; $Q_{eco average flow}$ is marked in green; $Q_{eco low flow}$ is marked in yellow. The months marked in orange are the breeding period for the potential dominant fish species related to the river typology where the cross-section is located.

. In step B2, in this case, c = 0.35 for the class of forecast “>100%”, c = 0.4 for the class of forecast “80–100%”, and c = 0.5 for the other classes of forecast.

The next step, B4, of the method starts with the identification of the breeding period for dominant fish species: trout, grayling and chub. The breeding period is March–June and October–December. For the months outside the breeding period, deviations/exceptions are allowed (due to the uncertainties of long-term forecasts) in order to reduce the values of ecological flow from $Q_{eco high flow}$ to $Q_{eco average flow}$ and from $Q_{eco average flow}$ to $Q_{eco low flow}$.The exceptions are not applicable in the first case study. Finally, the ecological flow values from Table 4 will be used in the real-time operation according to the forecasted hydrological regime of the area in which the cross-section is located (the Valea Cormaia River).

Similarly, as for the first case study, the other two case studies were performed. The ecological flow values which will be used in real-time operation according to the forecasted hydrological regime of the area where the cross-section is located are highlighted within

Table 5. The final values of the ecological flows related to the hydrological conditions forecast downstream of a hydropower plant water intake in the Teleajen River.

Class of Forecast	Jan.	Feb.	Mar.	Apr.	May	Jun.	Jul.	Aug.	Sep.	Oct.	Nov.	Dec.
>100%	1.03	1.03	1.42	2.43	2.43	2.43	2.43	1.03	1.03	1.03	1.03	1.03
80–100%	1.03	1.03	1.42	2.43	2.43	2.43	2.43	1.03	1.03	1.03	1.03	1.03
50–80%	1.03	1.03	1.03	2.43	2.43	1.42	1.42	1.03	1.03	1.03	1.03	1.03
30–50%	1.03	1.03	1.03	1.42	1.42	1.03	1.03	1.03	1.03	1.03	1.03	1.03
<30%	1.03	1.03	1.03	1.03	1.03	1.03	1.03	1.03	1.03	1.03	1.03	1.03

$Q_{eco high flow}$ is marked in blue; $Q_{eco average flow}$ is marked in green; $Q_{eco low flow}$ is marked in yellow. The months marked in orange are the breeding period for the potential dominant fish species related to the river typology where the cross-section is located.

and

Table 6. The final values of the ecological flows related to the hydrological conditions forecast downstream of an irrigation water intake in the Baboia River.

Class of Forecast	Jan.	Feb.	Mar.	Apr.	May	Jun.	Jul.	Aug.	Sep.	Oct.	Nov.	Dec.
>100%	0.152	0.152	0.180	0.180	0.180	0.152	0.123	0.123	0.123	0.123	0.123	0.123
80–100%	0.123	0.152	0.180	0.180	0.152	0.152	0.123	0.123	0.123	0.123	0.123	0.123
50–80%	0.123	0.123	0.152	0.123	0.123	0.123	0.123	0.123	0.123	0.123	0.123	0.123
30–50%	0.123	0.123	0.123	0.123	0.123	0.123	0.123	0.123	0.123	0.123	0.123	0.123
<30%	0.123	0.123	0.123	0.123	0.123	0.123	0.123	0.123	0.123	0.123	0.123	0.123

$Q_{eco high flow}$ is marked in blue; $Q_{eco average flow}$ is marked in green; $Q_{eco low flow}$ is marked in yellow. The months marked in orange are the breeding period for the potential dominant fish species related to the river typology where the cross-section is located.

.

4.3. Results of C—Real-Time Operation Component

Aiming to assure the real-time flow dynamics on the river downstream of the analyzed cross-section, the appropriate value should be selected, e.g., in case study 1, from the Table 4, month by month, according to the hydrological forecast. Therefore, hereafter, an example for case study 1 is shown, for the hydrological forecast of April 2021. One looks for the April forecast in the website of the N.I.H.W.M., Bucharest, Romania [17] (see Figure 6). The analyzed cross-section is located at the “50–80%” class of forecast. The corresponding ecological flow value for the “50–80%” class (see Table 4) for April is 0.995 m³/s ($Q_{eco high flow}$). Therefore, this amount of water should be preserved/delivered for aquatic ecosystem protection during the whole month of April.

The monthly hydrological forecast is performed at the beginning of each current month, considering the meteorological forecast for that month. The dynamics of flow required by the European documents is given by the real-time operation. The ecological flow (low, average, or high ecological flow) value is chosen each month, from Table 4 for the case study 1, according to the class of forecast, and the hydraulic structure should be operated accordingly during that month.

A similar procedure should be performed for the other two case studies during the real-time operation.

5. Discussion

5.1. Discussions on A—Quantity Component

The success of the method, meaning the achievement and preservation of good ecological status is assured by the calibration of the method, i.e., the way of choosing the locations for the calibration and the use of long series of data in natural flow regime. Therefore, natural water bodies with minimum anthropogenic impact were chosen, where fish species existed in the past and recent monitoring confirmed their existence over time. This confirmation attests that fish species have been preserved over time in the local conditions, with low, average and high flows corresponding to the natural flow regime.

The coefficients of the method link the hydrological regime with the elements of the aquatic fauna habitat (coefficients: river typological groups, dominant fish species and breeding periods). In addition, the validation of the hydrological optimum domain was carried out according to the optimum water depth for the dominant fish species.

The ecological flows in the computing cross-section were set according to local catchment conditions expressed, implicitly, by the hydrological optimum (computed using the monthly natural/naturalized flow) and by the river typological groups, dominant fish species, ecosystem characteristics, and biological activities (reproduction) and the period in which they occur. The hydrological parameters of the method (monthly average flows) can be derived in controlled river basins, as well as in uncontrolled catchments. Therefore, the RoEflow method is practical, robust, and easy to apply for a variety of hydrological, morphological and ecological conditions (the coefficients ${\beta }_1$ and ${\beta }_2$ are related to the river typology groups—mountain, hill and lowland—and are indirectly linked to altitudes, the mean rainfall, the mean air temperature, the lithological structure and potential fish fauna).

Using the natural hydrological regime opens the debate of using the natural regime for rivers regulated by hydraulic structures, as named in the conditions of article 4(3) of the Water Framework Directive, i.e., heavily modified water bodies (HMWB). However, the Guidance Document No. 31—Ecological flows, in the implementation of the Water Framework Directive, specifies on page 71:

…depending on the nature and severity of morphological alteration, the hydrological regime consistent with GEP (good ecological potential—the environmental objective of HMWB) may be very close to the ecological flows that would have been required in the same water body before its morphological modification.

The Quantity component assessment should be tackled in the same manner for rivers regulated by hydraulic structures, but when implementing the ecological flows, issues of technical feasibility or disproportionate costs may be considered, as mentioned in the Water Framework Directive. Therefore, further research should be carried out in order to compute the exact figure to be delivered downstream of a dam. However, this should be performed case by case. It should be mentioned that the main idea of the Water Framework Directive is the analysis of the deviations/changes from reference conditions (natural, non-anthropogenic conditions) and the application of some measures aiming at the restoration of the natural conditions, in which the benefits are greater than the economic losses.

Also related to the quantity component, another issue might be the impact of water pollution in the calculation of ecological flows. European legislation gives clear thresholds for the effluents discharged within the rivers, and requires measures for non-point pollution sources. The incorporation within the computing method for ecological flows of a certain water quantity to alleviate the impact of pollution might not be the right solution.

Another discussion may be the usage of the calendar year and not the hydrological year. This is the current Romanian practice. However, one might argue that, most probably, using the calendar year for the computation of the discharges used within the method, as shown above, the water quantity is higher, meaning that the aquatic organisms will benefit.

5.2. Discussions on B—The Dynamic Component and C—The Real-Time Operation Component

The use of the hydrological forecast when putting into operation the ecological flow has the advantage of considering the climatic factors of the moment, which can significantly influence the hydrological regime of the river. This approach might lead to the diminishment of the impact of implementing the ecological flows on the different water users (for example, in the case of the hydropower use, the reduction of the losses for power energy).

The use of monthly discharges for computing ecological flows is related to the reservoir operating rules and regulations for large dams.

Another important issue is that real-time operation assures a certain magnitude, frequency, duration, timing and rate of change of flow that play a role in shaping physical habitats, which influence the biotic composition and sustainability of aquatic ecosystems.

5.3. Discussions on the Whole Method

The case studies show that the mean annual final values of the ecological flows may vary in the range between 15 and 28% of the multiannual average discharge (depending on the forecasted hydrological regime). In the implementation phase, in the case of heavily modified water bodies, studies to justify technical infeasibility or disproportionate costs might be carried out in order to adapt the values of the ecological flow.

The limitations of the method are the monitoring of the implemented ecological flows and the detailed testing performed to demonstrate that the water quantity and dynamics assure consistency with a good ecological status according to the Water Framework Directive’s environmental objective for natural waterbodies.

The quantitative component of the RoEflow method can be updated accordingly when new information on the impact of hydro-morphological pressures on biological quality elements becomes available and experience from the ecological flow implementation and monitoring is gained. A further improvement might be achieved by linking the method to the species of conservation interest mentioned in the Habitat Directive.

In practice, monthly changing values of ecological flows that depend on actual forecasts might be a bit problematic in a long-term water management plan. Future research work will be carried out in this direction.

6. Conclusions

The paper describes a concept of assessing ecological flows in line with the conceptual definition of ecological flows provided in the European Guidance Document No. 31—Ecological flows in the implementation of the Water Framework Directive. The RoEflow method comprises three components: a Quantity component, a Dynamic component, and a Real-time operation component. The dynamic component relates the values of ecological flows (quantity of water) to the hydrological forecasting. The method of calculus is appropriate for real-time operation. There are three values (high, average and low water flow), which are relatively easy to be implement in the operating regulations of the hydraulic structures.

The concept/the ideas might be applicable in some European countries. It is a practical method.

The Quantity component comprises hydrological parameters (monthly average discharges) which are easy to be derive for large territories. The Romanian abiotic typological characterization of the rivers was based on the system B classification (Annex II of the Water Framework Directive) which is also used in other European Member States. The abiotic characterization is related to fish fauna, and is further linked with water quantity and with environmental objectives. The definition of the “hydrological optimum” as the domain ranging between $mi{n}_j\left\{mi{n}_i\left\{Q_{i,j}^{month. a}\right\}\right\}$ and $ma{x}_j\left\{mi{n}_i\left\{Q_{i,j}^{month. a}\right\}\right\}$, where: i is the month, j is the year, $Q_{i,j}^{month. a}$ is the monthly average discharge in year I, and month j computed in a natural/naturalized regime is an added value. Within this domain, the ecological flows should vary. This range for ecological flows may be useful in other European Member States using the system B classification. In addition, the range of values for the β coefficients may be useful. The value for each coefficient may be adapted for each site after detailed testing is performed.

The Dynamic component and Real-time operation component are important steps to link science (computing method) to practice (changing the operating rules of hydraulic structures for the new water demand, namely the environmental demand).

The ideas, to use the hydrological forecasting to ensure the water dynamics (required by the Water Framework Directive) and to develop the quantitative component considering putting the ecological flows into practice, might have importance for a broader audience.