Introduction

Eutrophication is mainly caused by the discharge of untreated domestic sewage, agricultural fertilizers, and industrial effluents into water bodies (Lobo et al. 2015). Eutrophication is a complex phenomenon among water managers, and researchers are reluctant to consider eutrophication quite distinct from the organic pollution in rivers according to the recent experiences (Kelly and Whitton 1998). New types of damage to aquatic environments, especially in rivers, have appeared, and it is clear that algae have many advantages as a bioindicator group for providing evidence of such damage. The Water Framework Directive (WFD) (European Communities 2000) requires assessing the ecological status with biological quality elements such as macrophyte and phytobenthos including algae in rivers. Over the last few decades, the assessment of organic pollution, eutrophication, and salinization has captured the attention of researchers and managers (Prygiel et al. 1999). Biological oxygen demand and ammonium are key indicators of organic pollution in lotic systems. Concentrations of these parameters increase as a result of discharges from waste water treatment plants, industrial effluents, and agricultural run-off. Phosphate and nitrate are used as indicators of eutrophication in flowing waters. Urban areas and industrial and agricultural areas might cause eutrophication. These factors can result in the growth of numerous plant species and therefore downgrade the ecological status (European Environment Agency 2020).

Salinization is also one of the important factors that affect diatom communities in rivers. Continental waters vary greatly in their mineral and ionic strength (salinity) mainly because of the geology and anthropogenic factors (Bere and Tundisi 2011). Kelly et al. (2019) suggested that conductivity reflects an interaction between geology and climate, with the hot summers of lowland Romania leading to evaporation that, in turn, increases dissolved ion concentrations. Therefore, conductivity plays also a major role in structuring the diatom assemblages. A major influence of geology on the selection of algal species and the assembly of benthic communities is the weathering of rocks and soils, and the consequent dissolution of chemical constituents, including essential nutrients for plant growth (Leland and Porter 2000). Potapova and Charles (2003) indicated that agricultural land use as an anthropogenic factor often increases the conductivity of river water, and these changes are reflected in algal communities. Salt leaching from irrigated soils can also elevate the naturally high salinity of many rivers. Stenger-Kovacs et al. (2018) investigated the functional characteristics of diatom taxa that alter considerably with conductivity, as proxy for natural and anthropogenic factors in the Carpathian basin. Saros and Fritz (2000) indicated that salinity may influence nutrient availability to primary producers, as well as nutrient requirements and uptake by diatoms. Therefore, eutrophication may have expanded modern diatom distributions in connection with salinity. Unfortunately, there is a lack of information on the responses of aquatic communities in rivers to concentrations of soluble salts commonly encountered (Leland et al. 2001).

Diatoms are often the most abundant and diverse group of algae within the phytobenthos and have been widely used for monitoring purposes (Kelly et al. 2008). Most Member States assumed that diatoms were the proxies for phytobenthos (Kelly et al. 2009). A number of diatom indices such as IPS (Pollution Sensitivity Index) (Cemagref 1982), IDG (Generic Diatom Index) (Rumeau and Coste 1988), IBD (Biological Diatom Index) (Coste et al. 2009) in France, EPI-D (Eutrophication Pollution Index-Diatoms) (Dell’Uomo 2004) in Italy, TDI (Trophic Diatom Index) (Kelly and Whitton 1995) in the UK, and TI (Trophic Index) (Rott et al. 1999) in Austria were developed for monitoring rivers of Europe. These indices have been used in many studies (Whitton et al. 1991; Ács et al. 2004; Blanco et al. 2008; Beltrami et al. 2012; Noga et al. 2016; Çetin and Demir 2019) for the classification of the ecological status.

Ecological status assessment studies based on diatoms in the Kızılırmak basin were rather limited in number. According to Akbulut et al. (2011), physical and chemical parameters with epilithic diatoms were investigated in ten stations in Kızılırmak and Delice rivers between May 2007 and September 2008. One hundred seventy three diatom taxa were identified in total, and the ecological status was classified by using diatom indices. EPI-D, IBD, IPS, IDG, TDI, and TI indices have quite a high correlation with physical and chemical parameters, and IDG, IPS, TDI, and IBD were the best diatom indices in the basin.

The aims of this study are (i) to test the performance of most widely used European diatom indices IPS, IDG, IBD, EPI-D, TDI, and TI according to the key indicators of organic pollution (biological oxygen demand and ammonium-nitrogen), eutrophication (orthophosphate and nitrate-nitrogen) and salinization (electrical conductivity) in the Kızılırmak basin and (ii) to evaluate the ecological status for phytobenthos with the best indicating diatom index for general degradation in the Kızılırmak basin according to the Water Framework Directive.

Materials and methods

Study area

The Kızılırmak basin is largely located to the east of Central Anatolia region between 37° 56′–41° 44′ north latitude and 32° 48′–38° 24′ east longitude. The Kızılırmak River is the longest river of Turkey (1.355 km) and discharges waters of an area of 82.181 km2 into the Black Sea (Yüce and Ercan 2015). Since the majority of the Kızılırmak basin is composed of gypsum rock series, its water is highly salty and bitter. Due to the lack of precipitation in summer period, the salt is not diluted, and salt levels in the Kızılırmak basin increase (Koç et al. 2018). In 2014, a total of 33 stations were surveyed seasonally (April, July, and October) in the Kızılırmak basin (Fig. 1). Two stations (R-13 and R-33) were dry in July. The coordinates and altitude information of the stations are given in Table 1.

Fig. 1
figure 1

Map of the study area

Full size image
Table 1 Coordinates and altitude information of the stations
Full size table

Sampling and analysis

Electrical conductivity (EC) was measured in situ, and water samples were stored in iceboxes and transferred immediately to the laboratory for later analysis. Orthophosphate (PO4-P), nitrate nitrogen (NO3-N), ammonium-nitrogen (NH4-N), and biological oxygen demand (BOD5) were measured according to APHA (2012).

The diatoms were collected by brushing submerged stones with a hard brush in riffle areas of rivers. In case there were no submerged stones at the station, epiphytic samples were taken by using a plastic bag (European Committee for Standardization 2014a). Samples were boiled with 30% H2O2 and HCl to remove organic matter. Samples were washed with distilled water in centrifuge (2500 RMP). The material was air-dried on cover glasses and mounted (European Committee for Standardization 2014b) with Naphrax® (Brunel Microscopes Ltd. Chippenham, UK). Light microscope (LM) observations were conducted using an OLYMPUS BX-51 (Mitsubishi UFJ Financial Group, Shinjuku, Tokyo, Japan) at the Dumlupınar University. For some small species (≤10 μm), scanning electron microscope (SEM) observations were made using a FEI NOVA NanoSEM 650 at Advanced Research Center (DPU-İLTEM) at the Dumlupınar University. The stubs were sputtered with 20 nm of gold using a Denton Vacuum Desk V (Phantom Export, Lisboa, Portugal). Diatoms were identified according to Krammer and Lange-Bertalot 1986, 1988, 1991a, 1991b; Reichardt (1999); Krammer (2000, 2002); Lange-Bertalot (2001); Lange-Bertalot et al. (2003, 2017); Werum and Lange-Bertalot (2004); Houk et al. (2010); Levkov et al. (2013); and Wojtal (2013). At least 400 valves were counted in each slide for all samples. Species having more than 10% of all counted valves in the diatom assemblage were defined as the most abundant.

Data analysis

Principal components analysis (PCA) was performed in XLSTAT (Addinsoft 2014) in order to reveal the general degradation in the sampling stations. Environmental variables were log-transformed in order to decrease skewness. IPS (Cemagref 1982), IDG (Rumeau and Coste 1988), IBD (Coste et al. 2009), EPI-D (Dell’Uomo 2004), TDI (Kelly and Whitton 1995), and TI (Rott et al. 1999) indices were calculated and transformed to the scale 0–20 by the OMNIDIA 6.0.8 (Lecointe et al. 1993). Linear regression analysis was performed by XLSTAT (Addinsoft 2014) between diatom indices and PCA’s first axis gradient of environmental variables for testing the performance of diatom indices in order to identify the general degradation of rivers. Pearson correlations between diatom indices and environmental variables were calculated with XLSTAT (Addinsoft 2014). Ecological status of the stations was classified according to the most correlated index with the general degradation. While blue color stations indicate high status in rivers, green, yellow, orange, and red stations indicate respectively good, moderate, poor, and bad status.

Results

In the research period, while the lowest mean EC (305.33 μS cm−1), NO3-N (0.04 mg L−1), and NH4-N (0.01 mg L−1) were measured in the station R-1, the highest mean PO4-P (4.46 mg L−1), BOD5 (74.55 mg L−1), and NH4-N (3.52 mg L−1) were measured in the station R-10. The station R-9 had the lowest PO4-P (0.04 mg L−1), and the station R-22 had the highest mean EC (7960.00 μS cm−1) (Table 2).

Table 2 Variation of environmental variables in stations by seasons (mean ± standard deviation)
Full size table

Two hundred fifteen diatom species were identified in total in the Kızılırmak basin. While the species as Achnanthidium minutissimum (Kützing) Czarnecki, Achnanthidium pyrenaicum (Hustedt) Kobayasi, Amphora pediculus (Kützing) Grunow, Cocconeis pediculus Ehrenberg, Cyclotella meneghiniana Kützing, Cymbella affinis Kützing, Diatoma moniliformis Kützing D.M. Williams, Navicula gregaria Donkin, Navicula lanceolata Ehrenberg, Navicula simulata Manguin, Navicula veneta Kützing, Nitzschia capitellata Hustedt, Nitzschia inconspicua Grunow, and Nitzschia palea (Kützing) W. Smith were the dominant species in the basin, Craticula accomoda (Hustedt) D.G. Mann, Encyonopsis minuta Krammer & E. Reichardt, E. subminuta Krammer & E. Reichardt, Luticola ventriconfusa Lange-Bertalot, Navicula cryptotenella Lange-Bertalot, Navicula cryptotenelloides Lange-Bertalot, Navicula phylleptosoma Lange-Bertalot, Navicula metareichardtiana Lange-Bertalot & Kusber, Navicula recens (Lange-Bertalot) Lange-Bertalot, Nitzschia amphibia Grunow, Nitzschia desertorum Hustedt, Nitzschia dissipata (Kützing) Rabenhorst, Nitzschia umbonata (Ehrenberg) Lange-Bertalot, and Ulnaria delicatissima (W. Smith) Aboal & P.C. Silva were the most abundant taxa in the stations (Fig. 2).

Fig. 2
figure 2

The most common taxa in the basin and abundant taxa in the stations 1, 2—Achnanthidium minutissimum, 3, 4—A. pyrenaicum, 5, 6—Amphora pediculus, 7—Cocconeis pediculus, 8—Craticula accomoda, 9—Cyclotella meneghiniana, 10—Cymbella excisa, 11—Diatoma moniliformis, 12, 13—Encyonopsis minuta, 14—E. subminuta, 15—Luticola ventriconfusa, 16—Navicula cryptotenella, 17—N. cryptotenelloides, 18—N. gregaria, 19—N. lanceolata, 20—N. metareicharditana, 21—N. phylleptosoma, 22—N. recens, 23—N. simulata, 24—N. veneta, 25—Nitzschia amphibia, 26—N. capitellata, 27—N. desertorum, 28—N. dissipata, 29, 30—N. inconspicua, 31—N. palea, 32—N. umbonata, 33—Ulnaria delicatissima. Scale bar 10 μm for LM images and 5 μm for SEM images

Full size image

PO4-P, NO3-N, BOD5, NH4-N, and EC variables were investigated in PCA analysis. The first two axes of PCA accounted for 78.43% of the total variance (Fig. 3). While the first axis (56.40% of the variance) of PCA indicated the organic pollution and eutrophication, the second axis (22.03% of the variance) of PCA indicated the salinization. Factor loadings were 0.87 (PO4-P), 0.84 (BOD5), 0.75 (NO3-N), and 0.72 (NH4-N) in the first axis and 0.79 (EC) in the second axis. The stations R-1, R-9, and R-4 indicated the lowest general degradation, and the stations R-18 and R-10 pointed to the highest general degradation in terms of environmental variables.

Fig. 3
figure 3

Principal component analysis for environmental variables in the stations

Full size image

Linear regression was performed between F1 scores (general degradation) of the stations in PCA analysis and diatom indices. The correlation between general degradation and diatom indices is shown in Fig. 4. While the EPI-D (R2 = 0.55), IPS (R2 = 0.53), IBD (R2 = 0.52) and TI (R2 = 0.50) were the best correlated indices with general degradation, the IDG (R2 = 0.44) and TDI (R2 = 0.32) had the lowest correlations in the Kızılırmak basin. Pearson correlations between diatom indices and environmental variables are given in Table 3. The EPI-D index showed significant correlations (p = 0.05) with all environmental variables including EC. The IPS, IDG, IBD, TDI, and TI indices were not significantly correlated with EC.

Fig. 4
figure 4

Linear regression between general degradation and diatom indices

Full size image
Table 3 Pearson correlations between diatom indices and environmental variables (n = 33) (p < 0.05 are in italics)
Full size table

Average scores of diatom indices and ecological status based on diatom communities at the stations are given in Table 4 with the percentage information about the species pool which is used by diatom indices.

Table 4 Average scores of diatom indices and ecological status based on diatom communities at the stations (Blue color indicate high status, green, yellow, orange, and red color indicate good, moderate, poor, and bad status, respectively) (Percentages indicate the species pool used by diatom indices.)
Full size table

Discussion

Regarding the distribution of the taxa found in the Kızılırmak basin, Achnanthidium minutissimum, Amphora pediculus, Cocconeis pediculus, Cyclotella meneghiniana, Navicula lanceolata, N. veneta, Nitzschia amphibia, N. dissipata, and N. palea were commonly reported in Turkish inland waters (Solak et al. 2012). Similar to the these results, Cochero et al. (2017) stated that Navicula veneta, Nitzschia amphibia, and N. dissipata were indicator species in nutrient-rich rivers with salinity pulse treatments in the Pampean plains of Argentina. Nitzschia dissipata (abundance with 13.4% in the station R-2) generally indicates from medium to high trophic levels (Lange-Bertalot et al. 2017). Accordingly, the BOD5 value of the station R-2 (11.63 mg L−1) was slightly higher than the other stations. In general, Craticula accomoda, Cymbella affinis, Diatoma moniliformis, Encyonopsis minuta, E. subminuta, Navicula cryptotenella, N. cryptotenelloides, and N. metareichardtiana were also common taxa in Turkish freshwaters (Solak et al. 2012; Maraşlıoğlu and Gönülol 2019; Taşkın et al. 2019). Craticula accomoda (abundance with 12.5% in the station R-10) is found quite frequently in strongly saprobically impacted running waters in particular in the waste water of sewage works, up to polysaprobic zone, and it is present in lower numbers in β-mesosaprobic running waters as well as in lakes with high nutrient levels, and it does not occur in the oligotrophic, mesotrophic, and dystrophic freshwater habitats (Lange-Bertalot 2001). In this context, the station R-10 was one of the most polluted rivers in the basin. Nitzschia umbonata (abundance with 15.6% in the station R-14 and 10.5% in the station R-18) was rarely distributed in eu- to polytrophic, saprobically impacted freshwater habitats, and less constantly recorded than N. palea or C. accomoda (Lange-Bertalot et al. 2017). Correspondingly, the stations R-14 and R-18 had poor ecological status in the basin. Nitzschia capitellata (abundance with 10.1% in the station R-14) was found very frequently and locally. It was also abundant from coastal waters to inland waters with medium electrolyte content, and also tolerant to the sewage load up to the polysaprobic level (Lange-Bertalot et al. 2017). The station R-14 was one of the stations with the highest electrical conductivity content in the basin with 2130 μS cm−1. Navicula recens (abundance with 10.7% in the station R-24) was found frequently also in brackish waters along coasts. In the remaining running water types and generally in lakes, it is found usually in very low numbers. Saprobity is tolerant up to (or slightly above) the β-α-mesosaprobic level (Lange-Bertalot et al. 2017). Therefore, the station R-24 had 2846.67 μS cm−1 conductivity and 27.93 mg L−1 BOD5. Luticola ventriconfusa (abundance with 10.2% in station R-27) is found widely distributed in moderately eutrophic rivers (Levkov et al. 2013). Accordingly, the station R-27 had moderate ecological status in this study.

Additionally, some taxa such as Navicula phylleptosoma showed characteristic features of electrolyte rich waters (abundance with 10.9% in the station R-22 and 10.2% in the station R-29), and N. phylleptosoma was described from coastal area of the Yugorsky-Region. However, the taxon exists in salt marshes around the Baltic Sea (Witkowski et al. 2000). In this context, the station R-22 had the highest conductivity values with 7960 μS cm−1. Navicula recens (abundance with 10.7% in the station R-24) and N. inconspicua (abundance with 20.7% in the station R-24) were especially abundant in some stations in the Inner Anatolian part of the basin (named as Salt Lake Basin) according to Akbulut (2010). Nitzschia recens was widespread species inhabiting brackish and freshwaters with high electrolyte content. Akbulut (2010) pointed out that salinity is one of the most important factors affecting and limiting the diatom composition. This also supports the presence of marine and brackish species in high conductivity rivers in this study. Encyonopsis subminuta (abundance with 25.2% in the station R-3) and E. minuta (abundance with 16% in the station R-12) were found in relatively high populations in Salda Lake with high conductivity in Burdur (Solak unpublished data). These species had been mostly found in the stations R-3 and R-12 with 2833.33 and 2032 μS cm−1 EC respectively as in this study.

PO4-P, NO3-N, NH4-N, and BOD5 variables indicate different pollutions in rivers. NH4-N and BOD5 both demonstrate the organic pollution load (European Environment Agency 2020). According to PCA’s factor loadings, the BOD5 (0.84) was more indicative than NH4-N (0.72) in the basin in terms of organic pollution. It can be explained by the lower indication of NH4-N with nitrification process in warmer climates. It could be converted to NO3-N much faster (Zheng et al. 2016). For instance, while the station R-1 (4.90 mg L−1) and the station R-33 (4.40 mg L−1) had similar BOD5 values, these stations had different NH4-N values in R-1 (0.01 mg L−1) and R-33 (0.05 mg L−1). However, the ecological status was classified in good status in both stations. The results showed that the BOD5 was more indicative than NH4-N in the stations.

NO3-N points to agricultural activities in the eutrophication process (European Environment Agency 2020). The importance of NO3-N might also depend on the relative concentrations of N and P in the system. Since the factor loading of PO4-P (0.87) is greater than NO3-N (0.75), it can be understood that phosphorus is a limiting element in the basin. For instance, the lowest PO4-P value in the station R-9 (0.04 mg L−1) showed the highest ecological status (EPI-D: 16.2) in the basin in terms of diatoms. According to the PCA plot, strong eutrophication and organic pollution were related with the salinization (electrical conductivity) in the basin.

Salinity is relevant in continental waters due to the geology and anthropogenic factors (Bere and Tundisi 2011). EC (factor loading with 0.79) was explained in the second PCA axis with 22.03% of the total variance following the eutrophication and organic pollution. Nevertheless, there is a huge interaction among geology, benthic communities, and dissolution of nutrients (Leland and Porter 2000). While the stations R-22, R-25, R-8, R-15, and R-24 with high EC values between 2846.67 and 7960 μS cm−1 are in bad and poor ecological status, the stations R-6, R-3, and R-5 also with high EC values between 2683.33 and 4183.33 μS cm−1 are in good ecological status (Tables 2 and 4). According to Stenger-Kovacs et al. (2018), aquatic ecosytems are exposed to naturally high conductivity. As it is a subject of criteria set by the WFD, there is an urgent need for developing a suitable quality index for their ecological status assessment. Most of the diatom indices are trained to indicate high salinity results as a result of human pollution (bad ecological status), and the reliability of these indices is highly questionable due to poor species pools. Therefore, the DISP (Diatom Index for Soda Pans) was developed using conductivity as a master environmental variable. It is crucial to carry out such index studies for rivers. Nevertheless, the distinction of good and bad ecological status in different stations with high conductivity was achieved with the high percentage species pool in the EPI-D index. Therefore, the general degradation should also be investigated in the basin.

In order to test the applicability of European diatom indices (IPS, IBD, TDI, CEE, and EPI-D), the Guadalquivir River catchment in Southern Spain was investigated by Martin et al. (2010). Similarly, they found that the EPI-D and secondarily the IPS provided the best results in the basin. Regarding to the Achnanthidium Kützing species, similar result was found in the study as well. The stations with a high percentage of the Achnanthidium genus, the EPI-D and IPS indices revealed quite different results because of the SEM technique used in this study for the identification of Achnanthidium species. For instance, the stations R-2 and R-7 where mostly the Achnanthidium species found were examined, and mean EPI-D values were 12.2 (R-2) and 10.8 (R-7) while mean IPS values were 14.4 and 13.0, respectively. It suggested that the identification of small species in Achnanthidium genus is essential, and it also affects the index results. According to Lange-Bertalot et al. (2017), Achnanthidium pyrenaicum occurs in oligotrophic and mesotrophic running waters. Bey and Ector (2013) also underlined that it was very sensitive to organic matter. Achnanthidium pyrenaicum was the most dominant species in station R-9 and the station had the highest ecological status in this study.

According to Dell’Uomo (1996), the EPI-D evaluates nutrient levels, organic pollution, and salt concentration in rivers. Correlations of the EPI-D values were highly significant with some related (BOD5, COD, nitrate, phosphate, chloride, and sulfate) or individual parameters (conductivity and total hardness) as mean values. The EPI-D is also related to the IPS index. Similarly, the EPI-D and IPS indices were the best correlated indices with general pollution in this study. Since the Kızılırmak basin has excessive salinity values, the EPI-D index was correlated significantly better than the IPS. Dell’Uomo and Torrisi (2011) tested also river Potenza with three new partial indices such as the TI-D (nutrients), SI-D (organic matter), and HI-D (mineral salts) by the broken down the EPI-D index in Italy. Both the EPI-D index and three partial indices (TI-D, SI-D, and HI-D) had high correlation with environmental variables. Coste et al. (2009) indicated that pH, conductivity, dissolved oxygen, biological oxygen demand, ammonium, orthophosphates, and nitrates were used for the establishment of the IBD index. However, the IBD index and conductivity were not significantly correlated. In this context, the EPI-D index was used to determine the ecological status of rivers in the Kızılırmak basin.

Conclusions

In this study, it is aimed to reveal the ecological status of the Kızılırmak Basin, where eutrophication, organic pollution, and salinization exist together, through diatoms by testing different diatom indices. The Water Framework Directive aims to maintain the good ecological status of all water bodies, and it puts more emphasis on the eutrophication/organic pollution in terms of the ecological status. Salinization is often disregarded since it is observed on a smaller scale and within limited areas in Europe. However, it is known that salinization has an influence on nutrient dynamics, and it needs to be evaluated together with the eutrophication and the organic pollution. A naturally occurring salinization problem is recognized in the Kızılırmak Basin due to its natural formation. Besides, eutrophication and organic pollution are also observed in the basin. Strong correlations were found between diatoms and physical and chemical parameters with the EPI-D index, which is among tested indices, and ecological status was classified on a sound basis. In addition, the EPI-D index is quite advantageous since it does not require a distinction among problematic Achnanthidium species, which occurs predominantly in the basin and which is difficult to identify. Thus, it becomes convenient in taxonomical studies. According to the EPI-D index, 73% of the stations in the basin are below good ecological status, and urgent actions need to be taken in line with the WFD. Only one station is in high status and can be regarded as a reference site. In accordance with the outcomes of the study carried out in the Kızılırmak Basin, it is suggested to use the EPI-D index in regions where eutrophication, organic pollution, and salinization are observed together.