1. Introduction
The concept of ecological or environmental flows was historically developed as a response to the degradation of aquatic ecosystems caused by human interventions (e.g., water overuse and flow regulations). The recognition of the need for a minimum amount of water to remain in a river for the benefit of characteristic fish species gave rise to terms such as minimum flows, instream flows and fish flows. A significant shift resulted in referring the concept to multiple river ecosystem aspects, recognizing the vital role of the entire natural flow regime in ecosystem structure and functioning.
The holistic approach to the so-called environmental flow assessment (EFA) in the 1990s was not just restricted to in-stream processes, but encompassed all aspects of a flowing water system, including floodplains, groundwater bodies, and downstream receiving waters such as wetlands, terminal lakes and estuaries. This approach also considered all facets of the flow regime (quantity, frequency, duration, timing, and rate of change), the dynamic nature of rivers and water quality aspects. Many interpretations of EFA exist. For instance, Tharme [
1] defined it as “an assessment of how much of the original flow of a river that should continue to flow down it and onto its floodplains in order to maintain specified, valued features of the ecosystem”. Today, the terms ecological or environmental flow, ecological reserve, environmental water allocation or requirement, environmental demand and compensation flow are equivalently used across different regions, and by different groups, to broadly define the water that is set aside or released to meet the environmental flow needs of water (eco) systems.
In the last decade, the link between river flows and livelihoods [
2] was considered by integrating the human dimension as part of a holistic approach to environmental flow assessment, covering issues such as aesthetics, social dependence on riverine ecosystems, economic costs and benefits, protection of important cultural features and recreation, and links to morphological processes. Today, the concept continues to evolve and is shifting from the traditional view of minimum water amounts to a more comprehensive and holistic understanding.
It is recognized that the health and sustainability of river ecosystems depends on multiple factors, including flow regime, river hydraulics (e.g., geometry of channel and riparian zone), level of exploitation, presence of physical barriers to connectivity, etc. [
3]. Thus, it involves a number of biological, geomorphological, physical, and chemical processes in a river that forms and maintains aquatic ecosystems [
4]. In addition, modern approaches (a brief summary of which is given in the next section) require the involvement of stakeholders and experts from multiple disciplines to also account for socioeconomic issues. However, an overall evaluation of all the above factors within an EFA study is extremely difficult. Despite the important methodological advances, providing sophisticated frameworks and modeling tools for quantifying the complex eco-hydrological and socio-economic processes and their interactions, the amount and time length of the required information, including extended field observations, is a key restricting factor. Therefore, in most real-world applications, the problem is often handled in the context of significantly limited data availability [
5,
6,
7,
8]. In such cases, it is essential to seek parsimonious EFA approaches, in terms of data, time, as well as expertise requirements.
In developing countries, the establishment of environmental flow policies is a major challenge, due to the conflicts arising between the preservation of the riverine environment and the need for large-scale infrastructure for water resource exploitation. An interesting case is the Kingdom of Lesotho, where an extended plan for water resources development (initiated in the mid-1980s) is in progress, known as the Lesotho Highlands Water Project (LHWP). A key component of this plan is the assessment of ecological flow requirements downstream of existing and planned dams. In this context, this study investigated the EFA problem at three pilot sites, located across the main watercourses of the country. Due to the aforementioned limitations, we developed a simplified version of the well-known DRIFT framework, taking advantage of hydrological and hydraulic data as well as ecological and fish-survey information. After this Introduction and a brief summary of literature approaches (
Section 1 and
Section 2, respectively), the article is organized as follows: In
Section 3 we provide a short description of the study area and data.
Section 4 and
Section 5, respectively, summarize the proposed methodology and its implementation. In
Section 6, we discuss the outcomes of our approach and its future perspectives.
2. Brief Review of EFA Approaches
As reported by Tharme [
1], early approaches for EFA already originate in the western USA, from the end of the 1940s. Today, numerous methods exist, of all levels of complexity. They can generally be classified into four categories: hydrological, hydraulic, habitat simulation and holistic. Comprehensive reviews of these methods are provided by Tharme [
1], Acreman and Dunbar [
3], Petts [
9] and, more recently, Acreman et al. [
10].
The hydrological approaches are the oldest ones and comprise a large variety of methods; from simple rules-of-thumb to more sophisticated procedures, all of which use streamflow data as the sole input, interpreted as the river “identity”. These should refer to hydrological conditions prior to anthropogenic disturbances. The task of extracting naturalized streamflow data from the original information, retrieved under human-modified conditions, is not straightforward [
7] and may also require the use of advanced modeling tools to accurately represent the pristine river system [
11].
In elementary hydrological approaches, also referred to as desktop or lookup-table methods, critical (i.e., minimum desirable) flow values are defined in terms of statistical indices [
3]. The oldest one, published in 1976, is attributed to Tennant [
12], who attempted to associate the streamflow regime of the wet and dry season, quantified as percentage of the mean annual flow (MAF), with the quality of fish fauna. This approach, also referred to as Montana method, was based on field surveys and systematic collection of width, average depth, and average velocity data. In European countries, a pioneering index-based approach was imposed by the French Freshwater Fishing Law of 1984, requiring that residual flows in bypassed sections of a river should be at least 1/40 of MAF for existing schemes and 1/10 of MAF for new ones [
13]. Similar standards have been employed in many countries and incorporated in the related environmental legislation. For instance, in Greece, the minimum average monthly flow has been generally used to determine the flow to be maintained below dams. In Spain, 10% of MAF is generally employed for river basins with limited information, while routine values in Portugal are 2.5–5.0% of MAF [
1].
The flow targets are also assessed by setting specific exceedance percentiles of flow duration curves, derived from statistical analysis of daily discharge records [
14]. For instance, the
Q95 (i.e., the flow which is equaled or exceeded 95% time) is adopted as a minimum standard in UK, Australia, Taiwan and Bulgaria, while Canada and Brazil typically use the
Q90 discharge. On the other hand, some countries consider much less conservative thresholds, such as the
Q364, which corresponds to the minimum daily flow of the year, and it is practically estimated as the 99.7% discharge quantile. The UK standard, i.e.,
Q95, was recently specified by a multidisciplinary team of leading water scientists and competent authorities, who are responsible for implementing the 2000/60/EC Water Framework Directive (WFD) in Europe [
10]. In this context, the team of fish ecologists recommended various abstraction thresholds as a percentage of flow encountered on the day, in excess of the natural
Q95, as the lower limit for fish maintenance. The
Q95 was also proposed by experts from multiple disciplines, on the basis of hydraulic data retrieved from 65 sites across the UK [
13].
Instead of imposing a time-constant flow constraint, more advanced hydrological approaches wish to preserve the variability of flows at multiple temporal scales (monthly, seasonal and annual) and thus they are supposed to be more ecologically relevant. The most representative ones are the Basic Flow Method (BFM) [
15], the Range of Variability Approach (RVA) [
16,
17] and its revisions [
18] as well as the Ecological Limits Of Hydrologic Alteration (ELOHA) framework [
19]. These methods attempt to describe the full flow regime by using a number of hydrological indices that are associated with ecological criteria. RVA introduces the so-called Environment Flow Components (EFCs), accounting for five different flow types, i.e., low flows, extreme low flows, high flow pulses, small floods, and large floods. This delineation of EFCs is based on the realization by research ecologists, that river hydrographs can be divided into a repeating set of hydrographic patterns that are ecologically relevant. It is the full spectrum of flow conditions represented by these five types of flow events that must be maintained to sustain riverine ecological integrity. Not only is it essential to maintain adequate flows during low flow periods, but also that higher flows, floods and extreme low flow conditions perform important ecological functions.
Hydraulic rating (also known as habitat retention) approaches assume that hydraulics is the key driver of the integrity of the river ecosystem. In this context, these approaches evaluate a number of hydraulic, morphological and geometrical characteristics (wetted perimeter, depth, velocity, etc.) and establish links with habitat availability of target biota. Input data includes historical flow records and cross-section data. The analysis is implemented at a few, carefully selected cross-sections, particularly in shallow areas (e.g., riffles) or areas with important ecological characteristics, which are considered as critically limiting biotopes. The obvious assumption is that the protection of the most critical hydraulic areas ensures the maintenance of the entire aquatic ecosystem.
Hydraulic rating methods can be viewed as preliminary appearances of habitat simulation or even holistic approaches [
1]. Since the available aquatic habitat, for given flow conditions, is mainly determined by the wetted perimeter of the channel, several methods have been proposed aimed at using the above geometrical characteristic as a basic tool for ecological evaluation. The key idea is that the wetted perimeter of shallow and wide rivers is more sensitive to flow changes, than narrow and deep ones [
3]. In this context, the wetted perimeter–discharge breakpoint has been extensively used to define optimum or minimum flows for fish rearing in the USA, from the mid-1970s. The breakpoint (also referred to as the inflection point) is the point where the slope of the stage–discharge curve changes (decreases), so that a large increase in flow results in a small increase in perimeter. The lowest breakpoint in the curve is taken to represent a critical discharge below which habitat conditions for aquatic organisms rapidly become unfavorable [
20,
21].
Nowadays, the use of hydrological indices or the wetted perimeter as unique indicators of the river health, and the means for determining ecological flows, is limited. In this respect, hydrological and hydraulic approaches are usually employed only in conjunction with process-based techniques that also account for water quality and biological criteria. For instance, habitat simulation methods assess the environmental flow requirements on the basis of detailed analyses of the quantity and suitability of instream physical habitat, particularly target species or assemblages (mainly fish), that are observed under different flow regimes. These use hydrological, hydraulic and biological data, at representative sites along the rivers, to represent the habitat conditions within hydraulic simulation tools, thus allowing for the establishment of a direct link between habitat and discharge. A well-known approach in this category is the instream flow incremental methodology (IFIM) [
22], which has been considered by some practitioners as the most scientifically and legally defensible methodology for assessing environmental flows [
1]. IFIM was designed to address a number of ecological components (including flow regime, physical habitat structure, water quality, and energy fluxes) in the context of management decisions; for this reason, it also comprises mechanisms for analyzing the institutional aspects of water resource issues. Typically, the minimum acceptable flow is established according to predictions of instream habitat availability that are matched against preferences of one or a few species of fish, using hydraulic and habitat rating methods, such as in-stream transect analysis and the Physical Habitat Simulation (PHABSIM) software, developed by the USGS [
23] (which is major component of IFIM).
Finally, the so-called holistic methods—a term introduced by Arthington et al. [
24]—are essentially decision-making processes that allow scientists and stakeholders from multiple disciplines to share data and knowledge [
25], thus indicating a shift from prescriptive to interactive approaches [
1]. Their rationale is based on the same philosophy, imposing that all major abiotic and biotic components constitute the river ecosystem to be managed and the full spectrum of flows, in terms of their temporal and spatial variability, constitute the flows to be managed. Obviously, these approaches are much more demanding in terms of data requirements as well as human and time resources, since their objective is to address the requirements of the entire riverine ecosystem, comprising instream and groundwater systems, floodplains and downstream receiving waters [
25]. In fact, they do not provide explicit ecological targets but rather focus on determining the effects of water resources management scenarios, thus being suitable for related studies and master plans. In the last two decades, a wide range of such methodologies has been developed and applied, initially in Australia and South Africa, and more recently in the UK. The most recognized ones are the building block method [
26,
27], the Downstream Response to Imposed Flow Transformations (DRIFT) framework [
24], and the benchmarking method [
2,
28].
DRIFT is an interactive EFA approach, developed and applied within several water resources development projects in Africa. Based on a comprehensive data-management process, it provides flow scenarios and descriptive summaries of their consequences, in terms of the condition of river ecosystems, for examination and comparison by decision makers. These are then used to determine desired future conditions and the relevant instream flow requirements. DRIFT is made up of four modules, i.e., biophysical, sociological, scenario development and economic. The first evaluates the present nature and functioning of the ecosystem, and provides predictions of how these will change under a range of different flow manipulations. The second module identifies subsistence users at risk from flow manipulations and quantifies their links with the river. In the third module, the outputs from the first two are brought together to produce the biophysical and subsistence scenarios, while the fourth module addresses the costs of mitigation and compensation.
As mentioned in the Introduction, the proposed approach is actually a simplification of the DRIFT framework, aimed at balancing the holistic but data (and time) demanding approach of the original method with the severe data limitations across the study area (i.e., Lesotho).
7. Conclusions
The main objective of this study was to provide an environmental flow policy at three pilot sites in Lesotho, located across the main watercourses of the country. Our methodology follows the rationale and concepts of state-of-the-art environmental flow assessment approaches, particularly the well-known DRIFT method, at the same time being more parsimonious and easier to implement. In this respect, it was adapted to local conditions using the available data (hydrological, hydraulic, and ecological) and abstract information of fish habitat preferences based on field surveys.
A new dynamic approach was proposed, linking several components of the riverine environment in conjunction with hydrological, hydraulic and ecological criteria and environmental restrictions, to provide a dynamic environmental flow policy that consists of:
A monthly-varying range of target flow releases that follows the physical regime, which can be seen as a reconstruction “mimic” of the total river environment restoration.
An artificial flooding schedule, by means of releasing extreme daily flows corresponding to return periods of two and five years, which are essential for maintaining crucial river elements, such as floodplain vegetation and sediment refreshment.
A novel conceptual approach for fish quality classification, which links water depths and associated flow thresholds with habitat characteristics, for a wide range of fish species.
The overall approach was employed at three representative river sections across Lesotho, draining catchments of different extent, and was next validated using long-term records of flow and sediment samples at two out of there pilot sites. Since different approaches provide (as expected) a wide range of target flow values, a synthesis of criteria is essential in establishing a proper EF policy. The major contribution of this research is the direct expression of a specific flow value with a fish quality index, which allows easy quantification of the impacts of the selected policy to the aquatic environment. The quite large range of target baseflow values per season, generally expressed in terms of Q90 (minimum) and Q60 (maximum), ensures flexibility for water managers, while occasional deviations are acceptable, through the concept of subsistence flows, estimated as the Q97.
An interesting conclusion of the combined analyses is that in most cases, the target value Q90 exceeds the so-called Min_Medium flow that corresponds to Fish Quality Class 2. Nevertheless, it is strongly recommended that the largest of the two flow values is employed. As shown in the case of Mabine station, the implementation of a stricter EF target, together with a carefully planned artificial flooding schedule, also favors the sedimentation processes, which is a very important component of the overall health of the downstream riverine system.
However, an even more beneficial EF policy is to assign, as a baseflow target, the critical flow that corresponds to Fish Quality Class 4, i.e., the Average_Min. Interestingly, at all examined stations, the latter is comparable with the lowest breakpoint on the wetted perimeter–discharge curve. This conclusion, which should obviously be validated in more studies worldwide, may imply that the two approaches, i.e., hydraulic and ecological, converge to similar outcomes. If this hypothesis is valid, although receiving reasonable criticism [
33], the problem of EF assessment, at least in a preliminary context, would be substantially simplified, since the breakpoint analysis is very simple, only requiring the use of easily-retrieved cross-section geometry data and the implementation of trivial hydraulic calculations.