1 Introduction

Wetlands are commonly expressed as nonpareil examples of ecosystems with notably high biological diversity. Such areas that provide facilities of feeding, living and fertility for a diverse set of organisms and creatures are considered to be natural assets not only for countries they are located in, but also for the entire world. Wetlands, which have primary roles in daily economic processes of the populations in the neighbourhood, hold a distinguished position inside all ecosystems with special regard to the protection and sustainability of natural equilibrium and biological diversity.

Although there are different factors that act on the ecological state and functioning of a wetland system, probably the most significant of all is the water flow entering the ecosystem. Increases in settled populations and correspondingly in demands for water make it difficult throughout the world to manage water and balance water allocations between aquatic ecosystems and other competing economic sectors such as agriculture, industry and hydropower generation. However, increased awareness on the need for including the environment as a separate sector in this mathematics looks quite promising with the resulting motivation on water managers in conducting detailed analyses, taking proper actions, building scenarios in an uncertain future and finally developing/implementing water management strategies with thorough consideration for a fair share of water. Right at this point, it is crucial to make predictions on environmental water requirements against existing and future capacities, perform assessments and take proactive actions intended for sustainable ecosystem management.

While there are various definitions in international hydrology literature, flows that contribute to the sustainability of aquatic habitats and water dependent ecosystem processes are often called environmental flows. Different approaches are outlined for determining environmental flows in studies which are performed in a wide spatial extent worldwide and consider the environment among the primary sectors expecting sufficient and good quality water. Belmar et al. (2012) investigated the ability to discriminate the natural hydrologic variation of streams in the Segura River Basin in Spain by two environmental classifications and the potential of using environmental classification scheme in the European Water Framework Directive (so-called WFD-ecotypes) for evaluating environmental flows. Meijer et al. (2012) discussed a new modeling framework on the RIBASIM (River Basin Simulation) model for determining environmental flow requirements in a more realistic and direct way for a fictitious case-study selected in Iran. Kim and Singh (2014) outlined an entropy-based hydrologic alteration assessment method by using gauged flow data to guide policy-makers in allocating water between different uses and mitigating anthropogenic impacts on natural flow regimes to have sustainable development. Mackay et al. (2014) studied flow metrics within the ELOHA (Ecological Limits of Hydrologic Alteration) framework for characterizing flow regime alteration expectedly resulting from dams and water management practices in southeast Queensland and then diagnosing hydrologic alteration in the study area to support the development of hydro-ecological relationships.

As briefly exemplified above, environmental flow assessment methods that differ mainly by data requirements, addressed ecosystem types, computation length and prediction accuracy can be classified into four generic groups: (1) hydrologic methods (through the use of certain percentile estimates for flow or other statistical approaches for appraising hydrologic characteristics), (2) hydraulic rating methods (depending on simple hydraulic variables, such as wetted perimeter or maximum depth of flow), (3) habitat simulation methods (making use of hydraulic habitat-discharge relationships, and (4) holistic methods (including more expanded set of variables such as geomorphology, hydraulic habitat, water quality, riparian and aquatic vegetation, etc.) (Arthington et al. 2004). Hydrologic and hydraulic methods together can also be counted among the so-called desktop approaches as a category defined for classifying methods in international literature (Dyson et al. 2008). Today’s tendency is more towards collating different approaches together (especially considering hydrology and habitat response to altered conditions) or operating hierarchically. Despite this methodological variety, the selection of a proper methodology for estimating environmental flows, in terms of both the potential and the requirements, is not a straightforward or fully guided task due to difficulties in understanding the relationship between the flow quantity/quality and ecosystem processes. Other constraints that complicate the selection would definitely include limitations with usable data and lack of finance, expertise, etc. (King et al. 1999; Tharme 2003). It is very much clear that holistic approaches, which consider hydrologic and ecological systems together, have bigger strength in performing realistic estimations; however, their operation may become quite costly when data, time and financial requirements are considered (Acreman and Ferguson 2010). In today’s world, one may easily conclude that there is not a single method that addresses all necessary pillars for environmental flow assessments (Acreman and Dunbar 2004; Dyson et al. 2008).

This study investigates the environmental flow potential, which is basically expected to vary under natural change in hydrologic conditions of a river system as well as supplementary anthropogenic impacts due to the existence of a dam storage and regulation facility in the basin upstream, with regard to the maintenance of ecosystem functioning within the delta located at the river outlet. A novel approach that is based on the identification of existing potential in terms of environmental flows and the estimation of environmental flow requirements through hydrologic assessments with reference to previous conditions is described. Flow deficit/surplus is then quantified in terms of comparative flow statistics on the assumption that the habitat extent and thus ecosystem requirements for water in a nature reserve would inherit from the former periods. Additional impacts of man-made structures built on the way of flow are displayed similarly. An algorithm is utilized for decomposing flow time-series into three flow components. These are separately considered in modifying flows first for eliminating the dam impact to obtain near-pristine (or close-to-pristine) hydrologic conditions and then reproducing modified flows to represent the hydrologic character observed previously in the river system. For performing the latter, the degree of hydrologic alteration between the considered periods is first assessed through some relevant indicators that are mostly significant in terms of environmental flows. With the methodological framework presented here, it is aimed to guide water managers in making decisions for allocating water between various demand sectors as well as the environment for sustainable management by answering related policy questions like how much extra environmental water would be needed for compensating environmental flow deficits or in what quantities a decrease in wetland habitat would result in case of inaction (provided that habitat-discharge relationship is well defined), or on the contrary, in what way a new water allocation scheme is structured if surplus is computed with reference to hydro-ecological conditions previously existing in the region.

2 Case Area

The study area is the delta of the Goksu River located inside the municipal borders of the city of Mersin along the Turkish Mediterranean coast (Fig. 1). The area was declared in 1990 as a Special Environmental Protection Zone for protection against pollution and exploitation and the conservation of natural resources and cultural assets in the area. The Goksu Delta is one of the few remaining areas in the world where sea turtles (caretta caretta, chelonias mydas) and blue crabs (callinectes sapidus) lay their eggs.

Fig. 1
figure 1

Goksu River and the delta

Full size image

The Delta is also regarded by the “BirdLife International” as a major refuge for more than 300 bird species. In Turkey, the area holds additional significance for being one of the primary reproduction areas for a wide range of fauna including reed rooster, summer duck, flamingo, heron, pelican, spurred pewit, long-legged marsh swallow, Izmir kingfisher, bee eater, moustached reed nightingale and the white throat warbler. As one of the best preserved wetlands in the world, the Delta was included in the list of the Ramsar Convention that indicates the wetlands of international importance.

3 Methodology and Computational Framework

The overall methodology covers analyses for unfolding environmental management needs in the wetland system, deciding upon a proper environmental management scheme, examining altered conditions with regards to stream flows received by the system and then estimating environmental flow requirements for maintaining the wetland ecology in an unthreatened state. It can be explained and further detailed in a stepwise approach as: (1) investigating the emergence of and quantifying the degree of alterations in flow conditions due to operations of existing water impounding structures, (2) providing data homogeneity between the pre- and post-impoundment periods (pre-damming and post-damming) to generate daily time series of near-pristine flows, (3) comparing hydrologic conditions especially in terms of low flow indicators between two successive historical data periods of comparable length (pre- and post-periods of 1953–1975 and 1976–2000), (4) adapting the post-period flows to the hydrologic conditions of the pre-period and re-generating the post-period flows with an imitated character, (5) estimating the needs towards a proper environmental management and assigning the type of management for protecting and rehabilitating the wetland, (6) predicting environmental flow time series, analyzing hydrologic changes with regards to different components of environmental flows and quantifying environmental flow deficits/surpluses based on inter-periodic comparisons, and (7) assessing added impact of the water impoundment in the catchment on supply of environmental flows deemed necessary for sustainable wetland management.

Investigation of alterations in flow regimes/characteristics due to either changing natural water conditions or human-induced impacts (e.g., damming and regulated water release), as introduced in points (1) and (3) of the methodology outlined above, is of considerable significance in environmental flow assessments. This is required at least for quantitatively unfolding impacts of changing conditions on river flows due to natural reasons (e.g., land cover changes after natural processes, changes in meteorological variables, also considering the climate change) and/or anthropogenic interventions (e.g., changing land uses mainly for economic considerations, commissioning of water impoundment structures, etc.). These would in return bring in expected rehabilitation or deterioration effects on habitat budgets and ecosystem resources. Indeed, environmental flow assessments do not necessarily require designation of pristine flow conditions in river systems, but they are greatly based on formerly existing conditions (potentially called near-pristine or close-to-pristine) prior to any quantified change, instead. To this end, Indicators of Hydrologic Alteration (IHA) software program (Kannan and Jeong 2011) was utilized in the study for assessing alterations in hydrologic conditions, concerning both natural and human-induced processes, in terms of the river flows received by the wetland system. In the IHA approach, significant changes in hydrologic conditions can be detected and quantified through a series of hydrologic attributes that are generated by comparing two successive periods before and after any known instance of intervention on water flows. The software mainly considers the range-of-variability (RVA) measure for designating significant hydrologic alterations (Richter et al. 1996, 1998) and computes 33 hydrologic parameters in five parameter groups each of which may have varying influences on ecosystem health. In the presented study, changes in hydrologic conditions were explored first between the pre- and post-impoundment periods of a hydraulic structure with known reservoir storage and regulation facilities, in order to detect added impacts on river hydrology and providing data homogeneity by eliminating such artificial impacts as far as possible. The task is then repeated between the early period 1976–2000 and the subsequent period 1991–2000 for explaining alterations in river and environmental flows due to natural climate behaviour.

Having detected any significant hydrologic alteration in time series data, the next step is the resolution of data compatibility problem between the compared periods with potentially different hydrologic characteristics. Data homogeneity between pre- and post damming periods was accomplished by first decomposing flow data series from both periods into three components; seasonality, moving average trend and residual (or remainder) series, and then conforming the data from the subsequent period to the previous through certain modification over the original data. In the study, data decomposition was achieved by following the Loess approach that is normally a seasonal trend decomposition technique based on a locally-weighted regression smoothing (Cleveland et al. 1990; Cleveland 1993). In this intermediate step, it was aimed first to perform seasonality adjustment for the post-impoundment period under the damming and regulatory flow impacts as all storage dams alter seasonal patterns to some extent. This is simply implemented by replacing the seasonality component of altered flows with the original prior to the commissioning of the dam. It is then necessary to conform the moving average trend component of flows to that of the pre-damming conditions. In the study, this was achieved by using flow-duration curves (FDCs), which are separately computed for both period flows, so that average flow values are re-computed for the latter period by interpolating new data values that correspond to the same probabilities of exceedance obtained from the FDC of the former period. The modified flows for the post-impoundment period were finally computed by re-combining seasonal (of pre-impoundment), moving average trend (adapted to pre-impoundment) and the original residual components for flow.

The next methodological step is the inspection of hydrologic alterations between the two periods assigned in the time windows of 1953–1975 and 1976–2000 from the entire length of the available flow gauging dataset. Upon sufficient conviction about altered hydrologic conditions between the periods, the need for conforming post-period flows to the pre-period conditions would emerge in this stage. This is necessary in order to generate a modified record of flows for the post-period that imitates the pre-period hydrologic character and then provide inter-periodic assessment. As these comparative assessments are expected to reflect relative conditions in terms of environmental flows, the next step becomes computing environmental flow time series. First, the flow data for the period 1976–2000 which were adapted to the pre-existing conditions and second, the original (yet already free from damming impacts) data series were considered here. To this end, environmental flow time series were generated through the use of monthly converted data series in GEFC (Global Environmental Flow Calculator) software platform developed by the International Water Management Institute (IWMI) (Smakhtin and Eriyagama 2008). In computing environmental flows in GEFC through the FDC shift approach (Smakhtin and Anputhas 2006), there is a need for assigning a proper environmental management class as an analysis parameter that allows computation of environmental flows to serve for maintaining a river ecosystem in a desired state or for providing rehabilitation of damaged populations or narrowed habitats.

The estimation of environmental flow requirements from the river flow data which were adapted to the earlier conditions was followed by the computation of environmental flows to be actually supplied under the recent conditions. Using both outputs, some additional analyses were performed in the study for indicating surplus/deficit rates in environmental flows through computed summary statistics. The IHA software is able to distinguish five different components of river flow, extreme low flows, low flows, high-flow pulses, small floods and large floods, with an extra capacity embedded for developing environmental flow recommendations and supporting the determination of environmental flow needs. Thus, daily flows from the same compared datasets were also utilized in the software to calculate these fractional components that are mostly considered important for protecting river ecosystem health (Mathews and Richter 2007). Beside the generic indications from monthly environmental flow analyses in GEFC, similar comparative assessments were repeated in the study of daily flows for individually unfolding impacts of changing hydrology on five environmental flow components due to the change of natural climate and external anthropogenic interventions as well.

4 Application

By considering an apparent set of parameters for hydrologic change, the IHA approach was mainly employed for exploring and highlighting potential hydrologic alterations, (1) between the periods before and after the commissioning of the Gezende Dam in 1990, and then (2) between the period 1953–1975 as the reference and the subsequent period 1976–2000 selected for comparison. The former comparison for unfolding impoundment and regulation effects on flows was made by concerning two IHA indices, base flow index and low pulse count, which relate mostly to environmental flow quantification for the wetland. From a comparison based on the median base flow index, which is simply computed through the division of 7-day minimum flow by the mean flow for year, one can easily say that there is explicit decrease in base flows in the post-impoundment period (Fig. 2). The three hydrologic alteration (HA) factors on the graph correspond to percentile groups of the lowest third, middle third and highest third. They indicate increased frequency of low flows with the low HA value of 1.467 and decreased frequencies for comparably higher quantities of base flows associated with negatively signed values for high and middle HA scores. The other index again used for comparing low flows displays a very relevant pattern with an increased median of low pulse counts between the periods and a positively computed high HA value that indicates increased number of low flow pulses in the post-period compared to the pre-impoundment reference.

Fig. 2
figure 2

Hydrologic alteration assessed between the periods 1953-1989 and 1991-2000 through a base flow and b low pulse count indices

Full size image

Concerning the alteration impact of the Gezende Dam justified in the previous step, post-impoundment period flows were modified for providing data homogeneity by eliminating added effects from the reservoir operation. For doing so, river flows in the two periods prior to and after the dam operations were considered in terms of seasonal flows, moving average flow trend and the residuals not explained by the other two (Fig. 3). The seasonality difference between the pre- and post-periods was detected to be limited without any remarkable shift of the yearly seasonal pattern except the seasonality contribution of negative values, during summer and autumn months, comparably lessened in the latter period. Moving average trend flows distinguished from the deseasonalized series in both periods demonstrated the estimated impact of reservoir operations in a way to decrease average flows and slightly decrease the variation range of flow discharges by flattening the flow duration curve resulted. In order to eliminate the added impact of reservoir operations on flows, post-damming flows were regenerated by considering the exceedance probabilities for individual flow data points, but embedded into the flow duration curve for the pre-damming period for computing adapted moving average flow data with the same probabilities of exceedance (Fig. 4). This modification constituted average flows increased in general terms as expected.

Fig. 3
figure 3

Daily flow time series and seasonal, moving average trend and residual components for the flows observed in the periods a 1953-1989 and b 1991-2000, prior to and after the dam commissioning

Full size image
Fig. 4
figure 4

a FDCs for the moving average trend identified in the pre- (solid line) and post-impoundment (dashed line) periods, b moving average flow component, adapted to the pre-impoundment conditions (solid line) against the original in the post-impoundment period (dashed line)

Full size image

The final step in efforts for eliminating the damming impact from the flows to adapt to the pre-damming conditions and thus generate a homogeneous dataset in the entire period was to re-calculate flows for the period 1991–2000 by combining the fragmented components. The increased pattern of the low flows in the adapted time series against the original series of the post-damming period in Fig. 5 is in agreement with the anticipated lowering impact of damming on low flows. It is visible that low flow quantities are particularly increased after the modification alongside other changes observed entirely in the flow time series.

Fig. 5
figure 5

Modified daily flow time-series displayed over the original daily series in the post-impoundment period

Full size image

In the presented study, analysis of environmental flows needed for maintaining wetland habitats required a relative assessment. Here, the environmental flow potentials from the recent past period (assigned in the study between the years 1976 and 2000) were compared to the target conditions that are desired to be retained from the earlier times (i.e., the reference period 1953–1975) in order to sustain ecosystem resources and habitat reserves in unaltered natural state. This analysis was conducted with the use of flow time series data that were homogenised in the previous step and by dividing the entire length of data into two periods of almost equal length. Here, the period 1953–1975 was considered to be the reference so that the hydrology potentials that prevail in the subsequent period 1976–2000 are compared to the reference state in this respect. The validity of hydrologic alteration between the two consecutive periods was investigated again with the consideration of the same two indices relating to low flow conditions as was implemented previously between the pre- and post-damming periods. Computed values of both indices pointed out a relative decrease in base flows in more quantitative terms, but also a decrease in the yearly average count of low flow pulses that accompanies reduction in base flows detected between the periods (Fig. 6).

Fig. 6
figure 6

Hydrologic alteration assessed between the periods 1953-1975 and 1976-2000 through a base flow and b low pulse count indices

Full size image

In the view of enough confidence on altered hydrologic conditions due to natural climate processes prevalent in the basin that drains to the wetland, some adjustments were additionally needed for the river flows recorded in the assigned post-period. This modification on flows allowed comparison of the original flows to the flows simulated in a way to imitate the reference hydrology conditions, which disappeared or were altered due to changing hydrology. Adjustment of flows based on antecedent hydrologic conditions was achieved through the flow defragmentation approach similarly as in the task of eliminating the impoundment effect of the Gezende Dam. In this respect, FDCs were extracted separately for the periods 1953–1975 and 1976–2000 from the moving average trend component, corresponding modification was performed on average flows, seasonality difference between the periods was removed and finally residuals were re-included for eventual computation of representative flow series. When FDCs from both periods are examined together, some flattening (or slope decrease) of the curve in the latter period around approximately the flow median reflects slightly decreased variability of flows and the corresponding figure of decreasing high flow discharges against certain increase in low flows on the contrary (Fig. 7). This altered character on the flows was basically simulated by modifying the flow discharges of the subsequent period to conform to the reference conditions of the previous period. The modified time series in the period 1976–2000 are shown in Fig. 8 with apparent reduction in low flows that mainly resulted from the simulation with a hydrologic character inherited from the reference conditions of the past.

Fig. 7
figure 7

FDCs for the moving average trend flows identified in the periods 1953-1975 (solid line) and 1976-2000 (dashed line)

Full size image
Fig. 8
figure 8

Modified and original daily flow time-series for the post-period 1976-2000

Full size image

The daily time series of both original flows in the period 1976–2000 and the same period discharges modified for adapting to the conditions of the reference period 1953–1975 were decomposed into environmental flow components in the IHA software to allow inter-periodic comparisons with respect to these flow components. Daily flow discharges of the post-period 1976–2000, again both original and the modified series, were then converted into monthly average flows to be used in comparative assessments on environmental flows computed by the GEFC software. Here, the environmental management class was a parameter required by GEFC in estimating environmental flows in accordance with the desired/targeted type of management. This selection was attributed to the study by Gul et al. (2014) that was previously conducted for a wide coverage of wetlands in Turkey. The according selection of class “C” provided to indicate a moderately modified ecosystem where “the habitats and dynamics of the biota have been disturbed, but basic ecosystem functions are still intact” (Smakhtin and Anputhas 2006).

5 Results and Discussion

From the environmental flow series generated monthly in GEFC using the original and modified series (Fig. 9), an initial assessment for the post-period 1976–2000 was conducted for grading the impact of changing natural climatic conditions on environmental flows and revealing corresponding deficit/surplus quantities through different statistical measures. This first assessment did not include the dam influence and additional impacts due to reservoir operations. The statistics indicate that altering hydrologic conditions between the two investigated periods created a positive impact on environmental flows with an increased potential by about 20 % with respect to the mean flow and about 12 % for the median environmental flow.

Fig. 9
figure 9

Monthly environmental flows generated in GEFC from the daily original (solely modified for damming impact) and modified (through adaptation to the pre-period flow conditions) series

Full size image

Furthermore, flow variability is also considered to be another factor that governs the need for environmental flows such that the rivers with more stable flow regimes may require relatively higher environmental low flows (so-called maintenance low flows) while rivers with more variable flow regimes, on the other hand, are expected to have relatively lower maintenance low flow requirements (Smakhtin and Anputhas 2006). Flow variability measured in the study through the statistics of standard deviation (std. dev.) and the coefficient of variation (CoV) seems to have experienced a slight decrease. This may eventually indicate a negative aspect for the wetland ecosystem and lead to requirements for higher quantities of maintenance low flows (Table 1).

Table 1 Descriptive statistics for the monthly environmental flows predicted from hydrologically-modified (1st column) and dam impact-isolated (2nd column) time-series of river flows in the post-period 1976–2000
Full size table

The positive appearance due to the increase observed in environmental flow potential under the computed natural change in hydrologic conditions is completely reversed in the post-damming period 1991–2000 when the storage and regulation effects of the Gezende Dam are taken into consideration (Fig. 10). Indeed, decreased environmental flow potentials, both in flow average and the median, are observed in this period by approx. 37 % (shown in the last column of Table 2) against the computed increases (by about 10 %) in flows of the same period isolated from the damming impact (given with the figures in the middle column of Table 2). Although it is not as severe as in the case of solely natural hydrologic change, decreased variability of flow still remains a negative signal for the ecological sustainability.

Fig. 10
figure 10

Monthly environmental flows generated in GEFC from the daily original (containing the damming impact) and modified (through adaptation to the pre-period flow conditions) series of the post-damming period

Full size image
Table 2 Descriptive statistics for the monthly environmental flows predicted from hydrologically-modified (1st column), dam impact-isolated (2nd column) and original (3rd column) time-series of river flows in the post-damming period
Full size table

Similar to the assessments for the monthly series investigated above, further detailing was provided for different fractional components that were assigned by the IHA software on daily time series with more targeted environmental flow concerns. For solely evaluating the impact of naturally changing climate without any count on storage and regulation actions, Table 3 shows the results from a comparative assessment conducted in the 1976–2000 period between the two different flow series. The first is the series adapted to the hydrologic conditions of reference period while the second is the original series where only the damming impact was eliminated. In the table, low flows are assessed through seasonal low flow medians while peak flow discharges are used for the remaining components. Spring and autumn low flows are associated with an increase by slightly bigger than 11 %, while winter periods seem to have experienced no considerable change in low flow quantities. The most critical season in terms of the change in low flows necessary for maintaining the ecosystem functionality in the wetland is indicated to be the summer with almost 9 % decrease in low flows arriving seasonally.

Table 3 Basic characteristics of environmental flow components (m3/s) determined in the post-period from the hydrologically adapted river flows (1st column) and the flows modified solely by considering the dam impact (2nd column)
Full size table

The assessments for environmental flow components separated from the daily flow series in the post-damming period 1991–2000 point at more critical problems. The relative assessment of the flow components separated from the daily flows with respect to the estimated natural change between the hydrologic conditions indicates no considerable change in terms of low flow appearance in the river, except the summer low flows reflecting a decrease rate again (from the comparison of the statistics in the first two columns in Table 4). When the expected impacts of dam storage and flow regulation are taken into account, however, bigger gaps, associated with sharp decreases of around 50 %, are observed in environmental flow components in the post-damming period from the comparison to the reference conditions of 1953–1975 (third column compared to the first in Table 4) which were selected in the presented study to reflect the pristine (or near-pristine) flow conditions.

Table 4 Basic characteristics of environmental flow components (m3/s) determined from the hydrologically adapted river flows (1st column), the flows modified solely by considering the dam impact (2nd column) and the original flows (3rd column) in the post-impoundment period
Full size table

6 Conclusions

Hydrologic methods were employed in the presented study for conducting desktop assessments on environmental flow quantities with reference to altering hydrologic conditions. They serve significant capacity for estimating associated impacts on habitat populations in ecological reserves and for assessing functional performance of related ecosystems, provided that habitat-discharge relationships are conveniently determined in later stages. These methods are one step ahead of the conventional basic techniques that implement environmental flow quantification through percentile estimation.

Holistic approaches, habitat modeling and even hydraulic rating methods would normally be superior to stand-alone hydrologic approaches. However, their widespread utility in a regional or country-wide study will be quite limited due to high demands for detailed data on topography, river geometry, habitat characteristics and species distribution, other than the river discharges. Although there is always the need for linking changes in environmental flow potential to habitat response through the definition of habitat-discharge relationships, e.g., from a habitat simulation model in terms of usable habitat area against the river discharge, this will be applicable only to aquatic habitats without mainly counting on other water-dependent terrestrial habitats (e.g., birds, mammals, reptilia, etc.).

Any gap in environmental flow quantities normally causes habitat contraction or loss in ecological functions. In this case, water managers may need to take decisions in balancing environmental flow releases (e.g., from a storage reservoir) to those that appeared under the antecedent hydrologic conditions for filling the gap. Information on deficit/surplus rates in environmental flow potential computed in comparison to the reference periods would remarkably serve to this end. However, a full compensation of environmental flow deficits would not be desirable in most cases due to parallel water stresses on other sectors competing for water. The post-impoundment period 1991–2000 in the presented case study for Goksu wetland provides a suitable example to this difficulty. In such cases, a sufficient comprehension of habitat-discharge relationship from a connected habitat simulation model would definitely help fairly balance the water supplies between the demand sectors (including the wetland environment).

The methodological framework followed in the presented study provides ex-post assessments for taking remedial actions towards habitat protection and ecosystem maintenance. Here, the assessment period should not be too long in a way to delay the fulfillment of compensation measures, nor too short in a way to limit statistical significance from a short sample. A periodic assessment scheme will be more convenient to this end. On the other hand, as river flow regimes are quite vulnerable to climate change, ecosystem health and functionality would definitely be subject to serious risks under substantially changing conditions (Piniewski et al. 2014). Hence, ex-ante assessments will definitely be of great use in maintaining vulnerable ecosystems before experiencing shortage in environmental flow quantities. Hydrologic models integrated with climate and land use change predictions would serve here for ex-ante assessments as such and thus help estimate future stresses on environmental flow potential for accommodating proactive measures.