Testing Macrophyte-Based Assessment Tools Developed Under the EU Water Framework Directive for Application in a Caucasus Region Country (Armenia)

Abstract

The UN framework of “Integrated Water Resources Management” recommends countries establish a sustainable management of their water resources, which relies on the availability of proper assessment indexes for the ecological status of surface waters. Such indexes have been developed by many EU countries under the EU Water Framework Directive. Non-EU countries may be interested in adapting such assessment tools. Hence, we describe here a test of three macrophyte-based assessment tools developed in EU countries for use in Armenia. All three indices were tested in a regionally adapted version where local species were entered into reference taxonomic lists, river types were assigned, and matching exercises of the results with benthic macroinvertebrates and physical and chemical parameters of the rivers were performed. The results show that the three tested assessment tools produced similar spatial patterns of ecological status changes, while the adapted version of the MTR index had the highest number of significant correlations with various metrics of macrozoobenthos and hydro-chemical parameters. We conclude that assessment scores for rivers based on macrophytes can be used for rivers of the Caucasus after regional adaptation of the reference macrophyte list. We recommend that the modified MTR index be introduced into the national hydrobiological system of Armenia after some additional adaptations.

1. Introduction

The UN framework of “Integrated Water Resources Management” (IWRM) recommends countries establish knowledge-based, systematic, and sustainable management of their water resources [1]. This is only possible if a proper monitoring system for the ecological status of surface waters is established [2].

The degree of integrity of surface waters can be assessed in terms of their hydrological, morphological, physical, chemical, and biological properties [3], and the European Union Water Framework Directive (EU WFD) [4] considers them all. However, to unveil in an integrative way the various human impacts that may occur during an extended time period at a scale of years, the focus is on biological indicators [5]. The EU WFD has adopted four biological quality elements (BQEs), benthic macroinvertebrates, fish, phytoplankton, and other flora (microphytobenthos and macrophytes), as bioindicators of surface water integrity. As a result, a variety of respective biological assessment tools have been developed based on them [6]. These four bioindicators differ in their sensitivity to the various human impacts on surface waters, such as pressures on hydrology, morphology, or water quality, but also show some overlaps in their sensitivity [7].

Macrophytes, including the periphyton (biofilm) growing on them, have several biological, ecological, and sanitary–hygienic functions [8]. Macrophytes provide habitat and food for fish and invertebrates [9,10,11], as well as several ecosystem services for humanity [12], such as food [13], flood protection [14], stabilization of water levels at low flow [15], nutrient retention and mineralization [16], detoxication [17], oxygen production, etc.

Generally, aquatic plants have a wide biogeographic distribution and partially span broad ecological conditions but mostly show clear growth optima at certain minimum or maximum levels of those environmental factors [18]. As a result of both physical habitat factors and inter-specific competition, macrophyte species form typical associations in response to environmental factors [19].

Macrophytes’ presence in streams influences the diversity, abundance, and taxonomic composition of other BQEs through the provision of habitat diversity and refuge [20]. However, different BQEs’ responses to various stressors are not the same [21,22,23]. Aquatic macrophytes mostly respond to artificially elevated concentrations of nitrogen and phosphorus but also to emissions of organic wastewater as well as alterations in hydrology and morphology [6,24]. Hence, macrophytes are the best biological indicators for the nutrient pollution of rivers together with benthic algae but need much more effort for taxonomic determination. Among aquatic macrophytes, hydrophytes (submerged macrophytes) usually represent better indicators compared to helophytes [25], as these are not directly influenced by habitat conditions in the river.

However, there are also some limitations for the use of macrophytes as bioindicators, as they are somewhat hampered by their “memory effect” and hence by the fact that the re-colonization of a restored river progresses relatively slowly. In consequence, aquatic macrophytes may often indicate the effects of river restoration efforts with a delay of several years [26]. Among river types, in mountain rivers, the abundance of macrophytes is usually low due to high hydraulic forces and frequent sediment transport during spates, which limits their use for monitoring routines there [27,28,29,30]. Besides a minimum abundance of macrophytes, another precondition for using macrophytes as bioindicators is the availability of a bioindicator species list adapted to the regional flora. The latter heavily constrains the use of aquatic flora in hydrobiological monitoring in non-EU countries such as countries in the Caucasus region [31]. Given these limitations, the use of benthic macroinvertebrates to assess the ecological status of surface waters is more common [6,30].

Armenia is one of the countries in the Caucasus region that strives to adapt the UN IWRM principles to a wider extent and heavily follows the EU WFD principles for its national monitoring routine. But because the study of microphytobenthos here is still in a very early stage [32], only the remaining three bioindicator groups of organisms are currently relevant. Benthic invertebrates are already used in the national hydrobiological monitoring routine [33,34], while European fish- and macrophyte-based indices still need to be adapted for use [31].

The aquatic macrophyte flora in Armenia has been well studied for the Hrazdan River Basin. Since the first studies back in the 1930s, 1940s, and 1950s [35,36] the taxonomic list for the basin has been regularly filled with new records [37].

The aim of this study was to test three existing aquatic macrophyte indices developed in EU countries for their applicability in the Caucasus region, as exemplified by the Hrazdan River.

The main criteria for determining applicability were the variability of assessment results under the gradient of impacts; the spatial pattern of changes in the results; compatibility with the results gained through the use of macrozoobenthos; and correlations with the hydro-chemical and hydro-physical parameters. Some discrepancies between the results derived through the use of macrophyte- and macrozoobenthos-based tools prove that they cannot be used in monitoring routines interchangeably but can be successfully used in parallel [38] for more comprehensive and reliable assessments.

2. Materials and Methods

Metrics and estimation of trophic status. Several macrophyte-based metrics are widely used for the bioindication of ecological status, e.g., macrophyte species areal cover estimates, abundance, species composition, and biomass. These metrics are also being incorporated into various assessment systems/indices. Here, we applied three macrophyte-based assessment tools, the Macrophyte Biologic Index for Rivers (Indice Biologique macrophytique en Rivière) (IBMR) [39], the Macrophyte Index for Rivers (MIR) [40], and the Mean Trophic Ranking (MTR) [41]. All these tools are designed to assess nutrient (N and P) enrichment, organic pollution, and channel degradation [42].

The IBMR system was established to replace the previous French system developed by [43] and is more targeted towards the requirements of the EU WFD. As this index was tested and validated for alkaline mountain rivers [39], we assumed it would be suitable for estimations of Armenian rivers, which are not alkaline, as well. The hypothesis was that geology does not significantly influence assessment systems in mountain regions. The index is calculated considering two features of each macrophyte taxon: (1) trophic class (CS_i; 1–5) and (2) ecological coefficient amplitude (E_i). It uses the following scale of cover (K_i): 1: <0.1%; 2: 0.1%–1%; 3: 1%–10%; 4: 10%–50%; 5: ≥50%. The IBMR score is calculated with the following formula:

$$IBMR={\displaystyle \frac{\sum _i{E}_i\ast K_i\ast {CS}_i}{\sum _i{E}_i\ast K_i}}$$

(1)

Table 1 sums up the trophic status classes according to scores of user assessment systems.

The MIR system was developed in Poland in accordance with the EU WFD principles based on the quantitative and qualitative evaluation of 153 indicator taxa. It can be calculated using the following equation [44]:

$$MIR={\displaystyle \frac{\sum _{i=1}^N{l}_i\ast w_i\ast p_i}{\sum _{i=1}^N{w}_i\ast p_i}}\times 10$$

(2)

where N is the number of species at the sampling site; l_i is the indicator value for the i-th taxon; w_i is the weighting factor for the i-th taxon; and p_i is the cover of the i-th taxon according to the nine-point scale developed by [41].

The range of MIR values are classified into five classes. However, following EU WFD requirements, the boundaries for those five classes are varied for different river types (Table 1). Because the intercalibration with the European river types has not been performed in Armenia yet [31], we checked if the selected sampling sites along the Hrazdan River possibly belonged to several stream types by carrying out hierarchical cluster analysis with IBM SPSS 17 software to gain the optimal clusters of sampling sites based on the composition of substrata. To validate the results of cluster analysis, the ANOVA test was performed. Based on the results, we provisionally assigned the sampling sites in different parts of the Hrazdan River to two river types, namely “A small upland river on a carbonate substrate (RW_wap)” and “Sandy lowland stream (PNp)”.

The British MTR system is based on presence and abundance metrics and uses a species trophic rank (STR (1–10)) scoring system according to the taxon tolerance to eutrophication. The final MTR score is calculated by summing up the scores of areal cover for STR species and then dividing it by the Species Cover Values (SCVs) of scored species. The overall result is multiplied by 10 to have a range from 10 to 100. The higher the score, the less eutrophicated the waterbody. In general, three classes of trophic status can be distinguished: more than 65—low probability of eutrophication; 25–65—some risk of eutrophication that requires further analysis depending on the local conditions; and less than 25—either heavily eutrophicated or affected by excessive load of organic matter and toxic materials or physically damaged [41].

Table 1. Trophic status classes according to IBMR and MIR systems [39,45].

IBMR Score	MIR Score (Type RW_wap)	MIR Score (Type PNp)	Trophic Status
>14	≥64.2	≥46.7	High (Very good)
>12 ≤14	≥49.7	≥36.8	Good
>10 ≤12	≥35.2	≥27	Moderate
>8 ≤10	≥23.7	≥16.3	Poor
≤8	<23.7	<16.3	Bad

Table 1. Trophic status classes according to IBMR and MIR systems [39,45].

IBMR Score	MIR Score (Type RW_wap)	MIR Score (Type PNp)	Trophic Status
>14	≥64.2	≥46.7	High (Very good)
>12 ≤14	≥49.7	≥36.8	Good
>10 ≤12	≥35.2	≥27	Moderate
>8 ≤10	≥23.7	≥16.3	Poor
≤8	<23.7	<16.3	Bad

The results obtained through macrophyte-based assessment tools were validated by (1) conducting Spearman correlation analysis among the results of ecological status assessment using macrophyte-based systems and hydro-chemical and hydro-physical parameters; (2) determining the concordance of results with macrozoobenthos-based biotic indices that reflect organic pollution and hydromorphological degradation well. The values of indices and metrics of macrozoobenthos were calculated with Asterics 4.04 software. Statistical analyses among the macrophyte-based indices and other parameters were performed using IBM SPSS 17 software.

Adaptation of used indices. To compile local reference taxonomic lists, we first identified missing species of macrophytes from the Hrazdan River in the existing reference taxonomic lists. For those species, we used the tolerance values from another list when possible. If the species was absent in all three reference lists for the used indices, assignment of tolerance scores was carried out based on expert judgement and considering species tolerance scores for species they form bidominant communities with. Also, we analyzed the ecological valence of the species based on their typical habitats and literature data before making a final decision on the score. Based on these principles, in some cases, we assigned the tolerance score of another species from the same genus, e.g., for Mentha longifolia, we used the score assigned to Mentha aquatica. Then, to understand whether adding new taxa into the reference lists would lead to significant changes in the results of assessments, we conducted a paired-samples t-test between the results of indices prior to and after adaptation as the Shapiro–Wilk test showed normal distribution for both non-modified (sig. from 0.054 to 0.997) and adapted (sig. from 0.093 to 0.282) index results. Concordance between the results was estimated by correlation analysis.

Study area and sampling sites. The Hrazdan River flows 141 km through the central part of Armenia and confluences with the Aras River at the border with Türkiye (Figure 1). It is the main water body in the Hrazdan River Basin Management Area (RBMA)—one of the six RBMAs in the territory of Armenia. According to EU-WFD-based river typology, it is a siliceous type [46], which makes it different from the alkaline upland rivers used in the elaboration of the IBMR and the carbonate-substrate-based small river type used in the MIR. The river is fed by snowmelt water, precipitation, groundwater sources, and Lake Sevan water [47]. More than 1.5 million people live in the catchment and leave their footprint via severe disruptions in river continuity and hydromorphological alterations [48], a flux of mineral water [49], discharge of industrial, agricultural, and household wastewater [50,51] and hydropeaking [52]. The river lacks wastewater treatment plants in general, and only one aeration plant is allocated in Yerevan (the lower course). Forests, grasslands, and shrublands prevail in the upper-course part, while built-up areas and agricultural land use prevail in the lower-course part (Figure 2). Macrophytes grow almost everywhere along the river course; thus, the Hrazdan River network forms a suitable testing area.

During the USSR period of Armenian history, a channel network was developed, involving heavy modification of the riverbed in the upper-course part to divert water from Lake Sevan for irrigation and the operation of seven hydroelectric power plants of the Hrazdan hydroelectric cascade. Additionally, the river continuity has been severely disrupted by the construction of three high dams, Marmarik, Aghbyurak, and Yerevanyan Lich, which all release hypolimnetic cold water in summer [46].

Thus, considering the multiple stress and accessibility issues in the mountain terrain, we established 12 sampling sites along the river.

Sampling and processing of materials. We conducted two field trips in May and July of 2021 to study macrophytes, while benthic macroinvertebrates were sampled only in July, when macrophytes are at the beginning of their vegetation process. At each station (

Table 2. Geographical description of sampling stations.

Sampling Station ^	Sampling Station Code	Description	Latitude	Longitude	Elevation (m.a.s.l.)
Geghamavan	1	Between Geghamavan and Tsaghkunq villages	40.5698	44.8906	1835
Aghbyurak	2	0.5 km downstream of Aghbyurak dam	40.4957	44.7360	1685
Solak	3	0.5 km downstream of Solak village	40.4608	44.6872	1519
Bjni	4	0.5 km upstream of Bjni village	40.4648	44.6737	1509
Arzakan	5	In the territory of Arzakan village	40.4347	44.6253	1460
Argel	6	0.5 km upstream of Argel HPP	40.3868	44.6033	1384
Getamej	7	0.5 km upstream of Getamej village	40.2829	44.5901	1225
Yerevan	8	0.5 km downstream of Yerevanyan Lich reservoir	40.1512	44.4600	881
Geghanist	9	1.5 km upstream of Aeratsia water treatment plant	40.1464	44.4363	871
Darbnik	10	4 km downstream of Aeratsia water treatment plant	40.1050	44.3799	836
Sis	11	0.5 km downstream of Sis village	40.0410	44.4096	829
Hovtashen	12	Between Hovtashen and Noramarg villages	40.0219	44.4431	829

Note: ^ Sampling was carried out between 15 and 21 May 2021 and between 6 and 8 July 2021.

), we surveyed a 100 m long area of the river channel and riparian zone for macrophytes and manually collected 240 plant samples in total. Samples were collected from 1 m² areas that were most abundant in macrophyte species, with seven replications. Species that were observed at the sampling site but not collected were also added to the taxonomic lists. Benthic macroinvertebrates were sampled in accordance with the EU standards [53,54] using a Surber sampler (frame is 0.09 m²; mesh size is 500 µm). Animals were separated from the substratum, grouped by families, and fixed in 96% solution of ethanol in situ. Further identification to the lowest possible level was performed in the laboratory under the Zeiss Stemi 305 stereo microscope (magnification 6×–40×) using several taxonomic keys, e.g., [55,56,57,58]. Given the high endemism and cryptism of the Caucasus fauna of invertebrates [59] and the lack of local taxonomic keys, we mostly identified genera.

At each sampling site, we measured temperature, velocity, substratum composition, channel width, and depth (Table 2). To calculate average velocity, we performed three measurements using a floating object—two along the banks and one at the central part where possible. To determine substrata, we randomly measured 50 samples’ maximum diameter [60] and then classified the substrata according to [61]. To measure average depth, we calculated the mean of all measurements along five transects made at 0.5 m intervals using a meter-stick or a marked rope with a weight at the end for the deeper parts.

For the abiotic parameters, we sampled five liters of water from each station and sent these for analyses to the laboratory of the Hydrometeorology and Monitoring Center, a state non-commercial organization (SNCO) of the Ministry of Environment of Armenia. In total, 40 parameters were analyzed, such as odor, transparency, pH, total mineralization, conductivity, BOD₅, COD, phosphates and total phosphorus, nitrite, nitrate, ammonium ions, sulfate, chloride, and hydrocarbonate ions, and 25 metals.

3. Results

3.1. Gaps in the Reference Lists

When applying the three assessment tools, it turned out that 17 out of 37 macrophyte species were missing in the reference list of at least one assessment tool (

Table 3. Macrophyte species recorded in the studied sampling sites of the Hrazdan River that are missing from the taxonomic lists of used indication tools.

Genus	Species	Indication Tool			# of Sites with Records
		IBMR	MTR	MIR
Nostoc	Nostoc sp. H		-	-	1
Chara	Chara sp. H		-		2
Hygroamblystegium	Hygroamblystegium tenax		-		1
Myriophyllum	Myriophyllum spicatumH	-			5
Phragmites	Phragmites australis			-	4
Glyceria	Glyceria fluitans		-		8
Persicaria	Persicaria hydropiper		-		7
Eleocharis	Eleocharis palustris	-			1
Scirpus	Scirpus microcarpus	-	-	-	3
Carex	Carex acuta	-			1
Carex	Carex riparia	-			1
Lythrum	Lythrum salicaria	-	-	-	3
Epilobium	Epilobium hirsutum	-	-	-	3
Bidens	Bidens tripartita	-	-	-	1
Mentha	Mentha longifolia	-	-	-	2
Mentha	Mentha aquatica		-		6
Juncus	Juncus inflexus	-	-	-	2
Total		10	12	8

Note: ^H—hydrophytes.

).

On average, three species out of all observed per study site were not included in the reference lists (Figure 3). Among them, Glyceria fluitans was recorded at eight stations and Persicaria hydropiper was recorded at seven stations, while several species were recorded only at one station (Table 3). Three of them are hydrophytes, while the remaining are helophytes. Most (32%) local species records were missing from the British MTR tool taxonomic list, while the MIR and IBMR systems did not have 21% and 27% of local species records, respectively.

3.2. Assignment of River Sections to River Types

The study sites were grouped into two clusters (Figure 4) characterized by stony versus fine particulate sediments (sand/mud/sludge). Stony substrata prevailed in eight stations, while the remaining four stations, namely Geghamavan (upper-course part), Darbnik, Sis, and Hovtashen (all in the lower-course part), were dominated by sand/mud or sludge (

Table 4. The basic hydro-physical and hydromorphological measurements. In each cell, the first figure represents the data for May and the second for July.

Sampling Station	Average Width (m)	Depth (m)	Velocity (m/sec)	Temperature (°C)	Mineral Substratum	Correspondence to Stream Type According to MIR System
Geghamavan	2/1	0.2/0.2	0.3/0.3	13/15	Sand and mud (80%), microlithal (10%), mesolithal (10%)	PNp
Aghbyurak	3/2	0.15/0.15	0.2/0.2	17/26	Mesolithal (40%), microlithal (30%), mud (20%)	RW_wap
Solak	6/6	0.3/0.3	0.9/0.9	13/15	Mesolithal (70%), microlithal (20%), macrolithal (10%)	RW_wap
Bjni	13/13	0.7/0.7	0.6/0.5	15/19	Sand and mud (40%), mesolithal (30%), macrolithal (30%)	RW_wap
Arzakan	12/12	0.5/0.5	0.4/0.4	15/19	Mesolithal (60%), microlithal (30%), sand, and mud (10%)	RW_wap
Argel	11/11	0.7/0.6	1.3/1	15/20	Megalithal (40%), macrolithal (35%), mesolithal (10%), sand (10%)	RW_wap
Getamej	11/10	0.5/0.4	0.4/0.3	12/17	Macrolithal (50%), mesolithal (35%), megalithal (10%), sand (5%)	RW_wap
Yerevan	15/15	0.5/0.5	1.1/1.1	13/21	Macrolithal (40%), mesolithal (20%), microlithal (20%), megalithal (10%), sand (10%)	RW_wap
Geghanist	12/12	0.5/0.5	0.9/0.9	13/19	Macrolithal (35%), mesolithal (25%), microlithal (25%), megalithal (10%), sand (5%)	RW_wap
Darbnik	20/20	0.6/0.6	0.3/0.3	24/19.5	Mud (20%), sludge (80%)	PNp
Sis	33/33	1.8/1.8	0.3/0.3	19/19.5	Sludge (50%), sand (30%), mud (20%)	PNp
Hovtashen	35/35	2/2	0.5/0.5	19/20	Sludge (60%), mud (20%), sand (20%)	PNp

).

The differences between the two clusters are statistically significant according to the ANOVA test (

Table 5. The result of ANOVA test for the mineral substratum categories.

	Cluster		Error
	Mean Square	df	Mean Square	df	F	Sig.
Fine particulates	10.212	1	0.079	10	129.643	0.000
Stony	10.059	1	0.094	10	106.879	0.000

Note: The difference is assumed to be significant at the p < 0.05 level.

).

The other hydromorphological parameters recorded at sampling sites, such as width, depth, flow velocity, and temperature, showed more random spatial distribution that mostly depends on artificial alterations of the channel, hydropeaking events, and water abstraction for irrigation and energy purposes.

3.3. Diversity of Macrophytes and Their Metrics

Among the 37 macrophyte species recorded along the course of the Hrazdan River, Cladophora glomerata, Glyceria fluitans, and Sparganium erectum species were the most common, while only a few records of Nostoc sp., Hygrohypnum ochraceum, Lemna gibba, Persicaria amphibia, Carex acuta, Carex riparia, Alisma plantago-aquatica, Bidens tripartite, Eleocharis palustris, Equisetum palustre, and Hygroamblystegium tenax species were made (

Table 6. Areal cover values of species recorded in the assessment of the trophic status of the Hrazdan River. For each species, the first value is the cover value according to the scale established for MTR [41], followed by the cover value according to the scale developed for IBMR [39].

N	Species	Sampling Sites
		1	2	3	4	5	6	7	8	9	10	11	12
1	Nostoc sp. H	-	-	6/4	-	-	-	-	-	-	-	-	-
2	Ulva intestinalis H	5/4	2/2	-	-	-	-	6/4	-	-	-	-	-
3	Cladophora glomerata H	5/4	3/3	2/2		2/2		5/3	6/4	3/3	2/2		2/2
4	Chara sp. H	-	-	7/4	3/3	-	-	-	-	-	-	-	-
5	Hygrohypnum ochraceum	-	-	4/3	-	-	-	-	-	-	-	-	-
6	Fontinalis antipyretica H	-	-	1/1	5/3	-	-	-	-	-	-	-	-
7	Stuckenia pectinata H	-	-	-	7/4	6/4	2/2	-	8/5	9/5	-	2/2	3/3
8	Zannichellia palustris H	-	-	-	5/3	7/4	3/3	2/2	-	-	-	-	-
9	Lemna minor H	-	2/2	-	2/2	2/2	2/2	-	2/2	-	1/1	-	2/2
10	Lemna gibba H	-	-	-	-	-	-	-	-	-	-	-	1/1
11	Lemna trisulca H	-	-	-	2/2	2/2	-	-	-	-	-	-	-
12	Ceratophyllum demersum H	2/2	-	-	-	-	-	-	-	-	-	2/2	3/3
13	Ranunculus trichophyllus H	2/2	-	-	6/4	5/3	5/4	2/2	-	-	-	-	-
14	Myriophyllum spicatum H	-	8/5	-	4/3	6/4	4/3	2/2	-	-	-	-	-
15	Phragmites australis	-	2/2	-	-	-	-	-	2/2	-	-	2/2	2/2
16	Glyceria fluitans	2/2	-	-	-	2/2	2/2	2/2	2/2	2/2	2/2	2/2	-
17	Catabrosa aquatica	2/2	2/2	-	-	2/2	-	2/2	-	2/2	-	-	-
18	Typha latifolia	3/3	-	-	-	-	-	2/2	-	-	-	2/2	2/2
19	Sparganium erectum	4/3	4/4	1/1	2/2	-	2/2	2/2	2/2	-	-	2/2	2/2
20	Persicaria hydropiper	2/3	-	2/2	-	-	-	-	2/2	2/2	1/1	2/2	2/2
21	Persicaria amphibia	-	-	3/3	-	-	-	-	-	-	-	-	-
22	Schoenoplectus tabernaemontani	3/3	3/4	-	-	-	-	-	5/4	-	2/2	-	2/2
23	Scirpus microcarpus	-	-	-	-	-	-	-	2/2	2/2	2/2	-	-
24	Carex vesicaria	8/3	2/2	-	-	2/2	-	-	2/2	-	2/1	-	-
25	Carex acuta	-	-	-	-	-	-	-	-	4/3	-	-	-
26	Carex riparia	-	-	-	-	-	-	-	-	-	1/2	-	-
27	Lythrum salicaria	-	-	-	-	2/2	2/2	-	2/2	-	-	-	-
28	Veronica anagallis-aquatica	2/2	-	4/3	2/2	-	1/1	6/4	-	-	-	-	-
29	Alisma plantago-aquatica	1/1	-	-	-	-	-	-	-	-	-	-	-
30	Epilobium hirsutum	-	5/3	-	-	-	2/2	-	3/3	-	-	-	-
31	Bidens tripartita	-	-	-	-	-	-	-	-	4/3	-	-	-
32	Mentha longifolia	2/2	-	3/2	-	-	-	-	-	-	-	-	-
33	Mentha aquatica	-	2/2	2/2	-	3/2	2/2	3/3	2/2	-	-	-	-
34	Juncus inflexus	-	2/2	-	-	-	2/2	-	-	-	-	-	-
35	Eleocharis palustris	3/2	-	-	-	-	-	-	-	-	-	-	-
36	Equisetum palustre	1/2	-	-	-	-	-	-	-	-	-	-	-
37	Hygroamblystegium tenax	-	-	2/2	-	-	-	-	-	-	-	-	-

Notes: Sampling sites: 1—Geghamavan; 2—Aghbyurak; 3—Solak; 4—Bjni; 5—Arzakan; 6—Argel; 7—Getamej; 8—Yerevan; 9—Geghanist; 10—Darbnik; 11—Sis; 12—Hovtashen. ^H—hydrophytes.

). In total, 13 were submerged macrophyte species, while the remaining 24 were emergent or partially submerged (helophytes), which are more vulnerable to grazing animals and bank modifications. The average areal cover was the highest for Stuckenia pectinata, Myriophyllum spicatum, Ranunculus trichophillus, and Cladophora glomerata species (Table 6). However, there were no species encountered in any of the sampling stretches.

3.4. Assessment of Trophic Status

The application of the three assessment tools, the IBMR, MTR, and MIR, on these macrophyte data collected in May and July of 2021 resulted in similar basic spatial patterns but clearly differed in the assessment of single sites (Figure 5). Adaptations in the reference taxonomic lists of indices allowed the trophic status to be re-calculated (Figure 5) on the basis of all species recorded at each station (

Table 7. Number of species used in the calculation of trophic status values by different indices.

Sampling Stations	N Species Used in Calculation of Each Index Value
	IBMR	MTR	MIR	Adapted Indices
Geghamavan	13	13	15	16
Aghbyurak	9	9	9	12
Solak	12	7	11	13
Bjni	9	9	10	10
Arzakan	10	9	11	12
Argel	8	8	9	11
Getamej	10	9	11	12
Yerevan	10	7	9	13
Geghanist	6	5	6	9
Darbnik	7	7	7	9
Sis	6	4	5	6
Hovtashen	10	9	9	10

). A comparison of the results derived from the initial assessment showed only slight changes in the trophic status in most of the stations, except Yerevan and Geghamavan for the IBMR, Yerevan for the MTR, and Geghamavan, Aghbyurak, and Darbnik for the MIR system.

We detected strong positive correlations between the results of all indices prior to and after the adaptation (

Table 8. Spearman correlation results among the scores of original and adapted versions of the three assessment tools studied (N = 12).

Indices	IBMR	MTR	MIR	Adapted IBMR	Adapted MTR	Adapted MIR
IBMR	1.000	0.754 **	0.606 *	0.878 **	0.765 **	0.730 **
MTR		1.000	0.946 **	0.715 **	0.928 **	0.949 **
MIR			1.000	0.616 *	0.862 **	0.935 **
Adapted IBMR				1.000	0.783 **	0.748 **
Adapted MTR					1.000	0.951 **
Adapted MIR						1.000

Notes: ** Correlation is significant at the 0.01 level (2-tailed). * Correlation is significant at the 0.05 level (2-tailed).

). Correlation coefficients were generally higher for the adapted versions of the indices. The strongest correlation was observed in both cases for the MTR-MIR, which is probably due to the partial similarity of the method of trophic score calculation.

We conducted a paired-samples t-test (

Table 9. The results of paired-samples t-test between the scores of original and modified indices. Scores all show a normal distribution according to a Shapiro–Wilk test of normality.

Pairs	t	df	Sig. (2-Tailed)
Adapted MIR—MIR (N = 12)	−1.526	11	0.155
Adapted IBMR—IBMR (N = 11)	−1.629	10	0.134
Adapted MTR—MTR (N = 11)	−5.127	10	0.000

Note: The difference is assumed to be significant at the p < 0.05 level.

) to reveal the significance of differences between the means of the scores of the original and adapted indices. The results show statistically significant differences only among the means of scores for the MTR index, where more adaptations were made.

3.5. Concordance of Results

Although in their initial state all three assessment tools showed almost similar patterns of trophic status changes along the course of the Hrazdan River, they frequently assigned a different trophic status for the same stations. In particular, the IBMR and MIR, which utilize a five-class division system, concluded different trophic statuses for 8 stations out of 12. The MIR assigned a one-class-higher status for six stations and a two-class-higher status for two stations, namely Darbnik and Sis. After the adaptation, differences in the results of the assessment were slightly mitigated. The adapted IBMR and adapted MIR concluded different trophic statuses for 7 stations out of 12, where a discrepancy was still observed between two classes at the Sis station. Both the Sis and Darbnik stations, where two class discrepancies in the results of the assessment were observed, were classified as PNp type by the MIR system. In both assessments, MTR showed fewer variations in trophic status among classes due to its three-category gradation. However, after modification, only Hovtashen station remained in the worst category.

3.6. Validation of Results

To validate the usefulness of macrophyte-based indices, we conducted correlation analysis with the macrozoobenthos-based metrics and the hydro-chemical and hydro-physical parameters. A total of 47 macrozoobenthos taxa (mostly genera) were recorded in the Hrazdan River during July of 2021 (

Table A1. Macrozoobenthos composition of the Hrazdan River sampling sites.

Genus	Sampling Stations
	1	2	3	4	5	6	7	8	9	10	11	12
Ancylus	-	1	62	4	-	-	-	-	1	-	-	-
Apatania	-	-	32	-	-	-	3	-	-	-	-	-
Asellus	-	-	-	9	-	-	-	-	-	-	1	2
Baetis	39	181	132	23	140	25	140	-	-	-	15	-
Bithynia	-	-	-	5	-	-	1	70	3	-	-	-
Caenis	8	-	-	-	9	-	2	-	-	-	-	-
Calopteryx	-	-	-	-	1	-	-	-	-	-	-	-
Chaetopteryx	-	-	-	-	-	12	-	-	-	-	-	-
Chironomidae gen.	-	208	208	5	8	0	15	105	108	-	-	136
Corixa	-	-	-	-	-	-	-	-	-	-	1	1
Dicranota	-	-	-	-	-	1	-	-	-	-	-	-
Dugesia	-	1	18	13	-	-	11	1	-	-	-	-
Dytiscidae gen.	-	2	2	1	-	-	-	-	-	-	-	-
Ecdyonurus	-	-	-	1	-	1	-	-	-	-	-	-
Elmidae gen. ad	-	-	13	-	2	-	-	-	-	-	-	-
Elmidae gen. lv	7	-	12	1	1	1	-	-	-	-	-	-
Ephemerella	2	-	20	9	4	18	-	-	-	-	-	-
Eristalinus	-	-	-	-	-	-	-	-	-	1	-	-
Erpobdella	1	3	3	31	1	-	2	-	-	-	-	-
Gammarus	16	18	1008	24	34	-	480	60	-	-	1	50
Glossosoma	-	-	-	1	1	3	-	-	-	-	-	-
Haemopis	-	-	-	1	-	-	-	-	-	-	-	-
Halesus	-	-	-	0	1	-	-	-	-	-	-	-
Helobdella	-	13	-	2	-	-	2	2	-	-	-	-
Hydrachna	-	-	1	2	-	5	1	-	-	-	-	-
Hydropsyche	-	-	6	2	1	22	-	-	-	-	-	-
Hydroptila	101	6	6	8	1	-	-	-	-	-	-	-
Laccophilus	-	-	-	-	-	-	-	-	-	1	-	1
Habroleptoides	-	2	-	-	-	-	-	-	-	-	-	-
Limnephilidae gen.	-	-	-	1	-	-	-	-	-	-	-	-
Limnephilus	-	-	1	1	-	-	-	-	-	-	-	-
Lumbriculus	-	-	2	-	-	-	-	-	-	-	-	-
Lymnaea	-	9	-	-	-	-	-	2	-	-	-	-
Muscidae gen.	-	-	-	1	-	-	-	-	-	-	-	-
Naididae gen.	-	24	-	1	2	1	-	-	-	-	-	-
Tubifex	-	-	-	-	-	-	-	7	12	-	-	37
Perla	-	-	-	1	-	-	-	-	-	-	-	-
Physa	-	-	-	5	-	-	-	-	-	-	-	-
Physella	-	-	-	-	-	-	-	-	1	-	2	-
Planorbis	-	1	1	-	-	-	-	-	2	-	-	-
Plectrocnemia	-	-	3	-	-	-	-	-	-	-	-	-
Radix	-	-	-	-	-	-	-	-	1	-	-	-
Rhyacophila	-	-	1	-	2	6	41	-	-	-	-	-
Sigara	-	-	-	-	-	-	-	-	-	-	1	-
Simulium	4	9	9	76	1	79	76	3	6	-	-	-
Syrphidae gen.	-	-	-	-	-	-	-	-	-	11	-	-
Tabanus	-	-	-	1	-	-	1	-	-	-	-	-

Note: Sampling sites: 1—Geghamavan; 2—Aghbyurak; 3—Solak; 4—Bjni; 5—Arzakan; 6—Argel; 7—Getamej; 8—Yerevan; 9—Geghanist; 10—Darbnik; 11—Sis; 12—Hovtashen.

). Based on taxonomic composition data, the Asterix tool returned 51 macrozoobenthos-based metrics. The Spearman correlation coefficient was significant (p < 0.05) between 27 macrozoobenthos-based metrics and at least 1 of the macrophyte-based indices (

Table A2. The Spearman correlation results between macrozoobenthos- and macrophyte-based indices and metrics.

Metrics		IBMR	Modified IBMR	MTR	Modified MTR_	MIR	Modified MIR
Number of taxa	Cor. Coeff.	0.402	0.519	0.539	0.734 **	0.454	0.582 *
	Sig. (2-tailed)	0.195	0.084	0.071	0.007	0.138	0.047
BMWP score	Cor. Coeff.	0.439	0.455	0.585 *	0.755 **	0.455	0.594 *
	Sig. (2-tailed)	0.154	0.138	0.046	0.005	0.137	0.042
- Ntaxa	Cor. Coeff.	0.424	0.519	0.534	0.737 **	0.424	0.586 *
	Sig. (2-tailed)	0.169	0.084	0.074	0.006	0.170	0.045
Average score per taxon	Cor. Coeff.	0.702 *	0.469	0.809 **	0.748 **	0.711 **	0.734 **
	Sig. (2-tailed)	0.011	0.124	0.001	0.005	0.010	0.007
BMWP score (Czech version)	Cor. Coeff.	0.432	0.434	0.579 *	0.753 **	0.463	0.581 *
	Sig. (2-tailed)	0.160	0.158	0.049	0.005	0.129	0.047
- Ntaxa	Cor. Coeff.	0.438	0.508	0.567	0.757 **	0.465	0.606 *
	Sig. (2-tailed)	0.155	0.092	0.055	0.004	0.128	0.037
Average score per taxon (Czech version)	Cor. Coeff.	0.516	0.371	0.578 *	0.699 *	0.497	0.587 *
	Sig. (2-tailed)	0.086	0.236	0.049	0.011	0.100	0.045
IBE aqem	Cor. Coeff.	0.527	0.567	0.547	0.766 **	0.394	0.575
	Sig. (2-tailed)	0.079	0.054	0.066	0.004	0.205	0.051
Diversity (Shannon–Wiener Index)	Cor. Coeff.	0.326	0.343	0.385	0.608 *	0.315	0.510
	Sig. (2-tailed)	0.301	0.276	0.216	0.036	0.318	0.090
Diversity (Margalef Index)	Cor. Coeff.	0.337	0.441	0.543	0.713 **	0.487	0.517
	Sig. (2-tailed)	0.284	0.152	0.068	0.009	0.108	0.085
Number of sensitive taxa (Austria)	Cor. Coeff.	0.397	0.378	0.747 **	0.824 **	0.708 *	0.757 **
	Sig. (2-tailed)	0.202	0.225	0.005	0.001	0.010	0.004
- Heteroptera (%)	Cor. Coeff.	−0.537	−0.640 *	−0.477	−0.570	−0.466	−0.640 *
	Sig. (2-tailed)	0.072	0.025	0.117	0.053	0.127	0.025
- Trichoptera (%)	Cor. Coeff.	0.649 *	0.509	0.648 *	0.712 **	0.520	0.657 *
	Sig. (2-tailed)	0.022	0.091	0.023	0.009	0.083	0.020
- Coleoptera (%)	Cor. Coeff.	0.461	0.278	0.745 **	0.569	0.693 *	0.683 *
	Sig. (2-tailed)	0.132	0.382	0.005	0.053	0.012	0.014
- Hydrachnidia (%)	Cor. Coeff.	0.622 *	0.628 *	0.600 *	0.599 *	0.508	0.553
	Sig. (2-tailed)	0.031	0.029	0.039	0.040	0.091	0.062
- EPT (%) (abundance classes)	Cor. Coeff.	0.541	0.313	0.547	0.580 *	0.496	0.481
	Sig. (2-tailed)	0.069	0.322	0.066	0.048	0.101	0.114
- EPT taxa	Cor. Coeff.	0.532	0.481	0.747 **	0.833 **	0.651 *	0.693 *
	Sig. (2-tailed)	0.075	0.113	0.005	0.001	0.022	0.012
- EPT taxa (%) (Austria)	Cor. Coeff.	0.598 *	0.439	0.670 *	0.735 **	0.586 *	0.635 *
	Sig. (2-tailed)	0.040	0.154	0.017	0.007	0.045	0.027
- EP taxa	Cor. Coeff.	0.428	0.389	0.628 *	0.770 **	0.511	0.621 *
	Sig. (2-tailed)	0.165	0.212	0.029	0.003	0.089	0.031
- EPTCBO (Eph., Ple., Tri., Col., Bivalv., Odo.)	Cor. Coeff.	0.570	0.491	0.804 **	0.851 **	0.676 *	0.726 **
	Sig. (2-tailed)	0.053	0.105	0.002	0.000	0.016	0.007
Number of families	Cor. Coeff.	0.415	0.497	0.537	0.729 **	0.432	0.581 *
	Sig. (2-tailed)	0.180	0.100	0.072	0.007	0.161	0.047
Number of genera	Cor. Coeff.	0.411	0.541	0.536	0.742 **	0.446	0.587 *
	Sig. (2-tailed)	0.184	0.070	0.072	0.006	0.146	0.045
- RETI	Cor. Coeff.	0.481	0.476	0.497	0.685 *	0.382	0.510
	Sig. (2-tailed)	0.114	0.118	0.100	0.014	0.221	0.090
- (%) gatherers/collectors	Cor. Coeff.	−0.505	−0.524	−0.466	−0.615 *	−0.431	−0.524
	Sig. (2-tailed)	0.094	0.080	0.127	0.033	0.162	0.080
- (%) shredders	Cor. Coeff.	−0.663 *	−0.395	−0.303	−0.174	−0.185	−0.267
	Sig. (2-tailed)	0.019	0.204	0.338	0.588	0.564	0.402
- Trichoptera_taxa	Cor. Coeff.	0.628 *	0.590 *	0.758 **	0.833 **	0.647 *	0.717 **
	Sig. (2-tailed)	0.029	0.044	0.004	0.001	0.023	0.009
Life Index	Cor. Coeff.	0.575	0.434	0.536	0.622 *	0.410	0.469
	Sig. (2-tailed)	0.050	0.159	0.073	0.031	0.186	0.124
Total N of significant correlations		6	3	13	24	7	18

Notes: ** Correlation is significant at the 0.01 level (2-tailed). * Correlation is significant at the 0.05 level (2-tailed).

) prior to or after the adaptation. Most of the correlations were positive. Macrophyte-based original assessment scores showed significant correlations with 14 macrozoobenthos-based metrics, while the adapted versions showed significant correlations with 26 metrics. After adaptation, the IBMR showed even fewer (only three) significant correlations with macrozoobenthos-based metrics. Neither the MIR nor its adapted version correlated with macrozoobenthos diversity metrics, and on average, the correlation coefficients for the adapted MIR were lower compared with those of the adapted MTR index. The number of significant correlations was the highest for the adapted MTR index (24 correlations), followed by the adapted MIR (18 correlations). Both the adapted MIR and adapted MTR index correlated well with the macrozoobenthos-based indices of taxonomic diversity and sensitivity widely used in Armenia, such as BMWP, ASPT, EPT, and number of taxa. This speaks to the validity of the adaptation results for macrophyte-based tools.

Spearman correlation analysis returned a statistically significant correlation with 18 hydro-chemical parameters out of 40 (

Table A3. The Spearman correlation results between macrophyte-based indices and those abiotic parameters which show significant correlation with at least one index.

Metrics		IBMR	Adapted IBMR	MTR	Adapted MTR	MIR	Adapted MIR
BOD5 mg/L	Cor. Coeff.	0.636	0.467	0.700 *	0.500	0.633	0.583
	Sig. (2-tailed)	0.066	0.205	0.036	0.170	0.067	0.099
Nitrite ion mg/L	Cor. Coeff.	−0.686 *	−0.683 *	−0.717 *	−0.850 **	−0.533	−0.683 *
	Sig. (2-tailed)	0.041	0.042	0.030	0.004	0.139	0.042
Ammonium ion mg/L	Cor. Coeff.	−0.385	−0.317	−0.533	−0.667 *	−0.450	−0.433
	Sig. (2-tailed)	0.306	0.406	0.139	0.050	0.224	0.244
Sulfate ion mg/L	Cor. Coeff.	−0.577	−0.467	−0.783 *	−0.817 **	−0.683 *	−0.667 *
	Sig. (2-tailed)	0.104	0.205	0.013	0.007	0.042	0.050
Chloride ion mg/L	Cor. Coeff.	−0.762 *	−0.733 *	−0.333	−0.500	−0.167	−0.267
	Sig. (2-tailed)	0.017	0.025	0.381	0.170	0.668	0.488
Nitrate ion mg/L	Cor. Coeff.	−0.720 *	−0.717 *	−0.483	−0.417	−0.317	−0.500
	Sig. (2-tailed)	0.029	0.030	0.187	0.265	0.406	0.170
Hydrocarbonate ion mg/L	Cor. Coeff.	0.502	0.576	0.780 *	0.881 **	0.644	0.831 **
	Sig. (2-tailed)	0.168	0.104	0.013	0.002	0.061	0.006
P mg/L	Cor. Coeff.	−0.603	−0.667 *	−0.250	−0.517	−0.050	−0.250
	Sig. (2-tailed)	0.086	0.050	0.516	0.154	0.898	0.516
Ca mg/L	Cor. Coeff.	−0.695 *	−0.767 *	−0.150	−0.250	−0.017	−0.200
	Sig. (2-tailed)	0.038	0.016	0.700	0.516	0.966	0.606
Ti mg/L	Cor. Coeff.	−0.603	−0.683 *	−0.067	−0.267	0.017	−0.100
	Sig. (2-tailed)	0.086	0.042	0.865	0.488	0.966	0.798
V mg/L	Cor. Coeff.	−0.628	−0.700 *	−0.667 *	−0.833 **	−0.500	−0.700 *
	Sig. (2-tailed)	0.070	0.036	0.050	0.005	0.170	0.036
Fe mg/L	Cor. Coeff.	−0.377	−0.367	−0.600	−0.767 *	−0.517	−0.550
	Sig. (2-tailed)	0.318	0.332	0.088	0.016	0.154	0.125
Mn mg/L	Cor. Coeff.	−0.385	−0.317	−0.567	−0.700 *	−0.450	−0.467
	Sig. (2-tailed)	0.306	0.406	0.112	0.036	0.224	0.205
Co mg/L	Cor. Coeff.	−0.628	−0.667 *	−0.417	−0.650	−0.233	−0.400
	Sig. (2-tailed)	0.070	0.050	0.265	0.058	0.546	0.286
Ni mg/L	Cor. Coeff.	−0.669 *	−0.717 *	−0.200	−0.400	−0.117	−0.200
	Sig. (2-tailed)	0.049	0.030	0.606	0.286	0.765	0.606
Cu mg/L	Cor. Coeff.	−0.594	−0.633	−0.633	−0.800 **	−0.467	−0.633
	Sig. (2-tailed)	0.092	0.067	0.067	0.010	0.205	0.067
Mo mg/L	Cor. Coeff.	−0.368	−0.233	−0.767 *	−0.700 *	−0.650	−0.650
	Sig. (2-tailed)	0.330	0.546	0.016	0.036	0.058	0.058
Pb mg/L	Cor. Coeff.	−0.343	−0.467	−0.467	−0.700 *	−0.400	−0.533
	Sig. (2-tailed)	0.366	0.205	0.205	0.036	0.286	0.139
Total N of significant correlations		5	9	6	10	1	4

Notes: * Correlation is significant at the 0.05 level (2-tailed).; ** Correlation is significant at the 0.01 level (2-tailed).

). The number of significant correlations increased by three for the adapted MIR and by four for the IBMR and MTR index. The lowest number of correlations was found in the case of the MIR, and the highest was found in the case of the adapted MTR index. After the adaptation, most of the indices showed a significant negative correlation with all trophic parameters. The adapted MTR index also showed significant negative correlations with trace elements, which opens new opportunities for its wider use to assess the ecological status of the rivers in Armenia. However, further detailed investigations in this regard can shed more light on this perspective. The current results show that all adapted indices performed better after the adaptation, with the adapted MTR better reflecting the changes in the components of the ecological status of the Hrazdan River. Thus, after the adaptation, macrophyte-based tools can be used in ecological status assessments of Armenian rivers, which requires some additional adaptation to the five-level gradation and recognition of precise thresholds of each status class.

4. Discussion

4.1. Limitations for the Use of EU WFD Macrophyte-Based Monitoring Tools in Mountain Regions Outside of the EU Countries

There are certain challenges with the use of macrophytes as bioindicators of the ecological state of mountain rivers, considering their presence/absence in a given river, growth, and spread due to interactions between several environmental factors, such as river depth, flow velocity, frequency of spates, water chemistry, nutrient concentrations, and light availability [62]. In the Caucasus region and Armenia in particular, many streams are lacking macrophyte diversity in the middle and upper stream parts [63] because of spates, high water velocity, heavy shadow from the forests, and lack of organic load because of the uneven distribution of the population [64]. Because most of the population is distributed at low altitudes and close to the rivers [65], the lower-course parts are experiencing multiple stress, which further limits the diversity of macrophytes and makes the macrophyte-based indices less reliable [66]. Our results have shown similar spatial patterns for the Hrazdan River in terms of both the diversity and areal cover of macrophytes (please see Table 6). Furthermore, mountain streams carry significant hydroenergetic potential, which is essential for the Caucasus countries [63], and the rivers here are mostly dammed for energy and irrigation needs [67], which disrupts river connectivity and influences macrophyte assemblages. Although macrophytes respond well to hydromorphological and general degradations [23], which are obvious at the Aghbyurak, Bjni, Getamej, and Yerevan stations, land grabs in the coastal zone and diversion to recreation zones or pastures poses further challenges for the use of macrophyte-based indices. Furthermore, the sampling period and shading of the site were found to be major factors affecting variability in macrophyte assessment results when using the MTR system in Europe [68], which, in the case of shading, is not relevant for the Hrazdan River but should be carefully considered for other rivers in the Caucasus. EU macrophyte-based assessment systems have further limitations for their use outside of the EU, as they rely on proper river typology and require intercalibration exercises for wider use [69]. Such challenges have not yet been overcome for the Caucasus counties, either [31]. Based on the derived results, we can conclude that geology and typology in general significantly affect the results of indices, thus reducing the applicability of EU-based macrophyte indices outside of the EU.

4.2. The Adaptation of EU-Based Indices for Armenia

The EU WFD creates opportunities for the member states to harmonize monitoring systems and adapt macrophyte-based indices for their use; however, the approaches are different [70,71]. We tested three different systems via adapting local species to the reference lists, assigning river typology, and carrying out statistical analyses of the results with benthic macroinvertebrates and the physical and chemical parameters of the rivers. The principle utilized to compile reference lists using local species worked well for MTR and the MIR, as the significance of changes and concordance with other parameters increased with the number of species added. Similar results were derived by [70] for Poland while adapting MTR system; however, at this stage, we prefer not to select a set of indicator species from the overall taxonomic list, as specific analyses were not broadly conducted for that. Also, [41] cautions that while using the MTR approach, the trophic statuses cannot be compared between sites that are significantly different in their hydromorphological parameters; however, it is hard to define which changes are significant to consider in the mountain rivers of the Caucasus, where such deliberate hydromorphological alterations can be found everywhere [31].

The principle of assigning existing river typology generally failed as an approach because adapted macrophyte-based systems having type-based classifications in ecological status assessment did not show significant changes in the mean values or concordance with other parameters. Discrepancies in the results of assessments using different macrophyte-based systems without proper adaptation are well known for other countries as well [42]. As a solution, the Reference Index was suggested for use in Hungary [42], with the further development of taxonomic lists and assignment of the tolerance scores for local species. Muratov et al. [71] conducted a similar adaptation of the MIR system for Kazakhstan based on the analysis of indicator species lists and the adaptation of the local list by adding local indicator species missing in the MIR. However, this still requires more in-depth studies for making its use feasible in Armenia.

4.3. The Reliability of Results

Considering that both the geology and altitude class of the Hrazdan River is different [46] from those of the regions the tested systems were developed for, the MIR and IBMR showed less reliable results at this stage in general. However, after adaptation, the MIR appeared more reliable for the upper-course parts of the Hrazdan River, as it was initially modified for upland river types. As [44] concluded, the IBMR correlates less with environmental parameters than the MIR, which is also true for mountain rivers in Armenia. Therefore, for Caucasus mountain rivers, separate systems should be developed for different river types with different lists of indicator species and their indicative values in accordance with the local environmental conditions, which we see as one of the next steps in ensuring the adherence of an Armenian hydrobiological monitoring system with the EU WFD requirements.

Although the MIR values for all stations were higher than the MTR score in their initial state, after adaptations, the MTR values were higher for some stations. Hence, missing species are influential for the overall assessment results in Armenia, as was shown previously for Hungary [42].

Because of its more general classification of trophy status and lower sensitivity to geology and altitude, MTR should work well for the areas with a heavy nutrient load, as was shown in the EU-based STAR project [72]. In the case of Armenia, MTR weakly reflects the abiotic parameters in the lower-course part of the Hrazdan River, where multiple stressors have impacted the river [73,74]. The diversity of macrophytes in this part, especially that of hydrophytes, is very low, and as suggested by [75], we tend to think that macrophytes should not be used for ecological status assessment for mountain rivers in such conditions.

As our results show for the IBMR, inconsistency between the alkaline and siliceous river types could be one of the major reasons why the French index does not work properly for the Hrazdan River even after the completion of reference taxonomic lists. Although the MIR showed better results after adaptation, further intercalibration exercises could improve the consistency of results for Armenian rivers. The gradient of impact from wastewater discharge among the stations [73] also varied widely, which reflects the cumulative effect of the population in upstream settlements [76]. Thus, the variation in the MTR and MIR within three classes and within one class for the IBMR proves the inappropriateness of the latter for use in a monitoring routine in Armenia at its current state.

Unlike the lowland rivers, freshwater discharge from the springs in mountain regions is leading to the development of near-untouched ecosystems, and the Solak station is an example of such an area, with the best results from all macrophyte-based indices. This, in general, proves that the EU macrophyte-based indices are applicable for mountain streams, which was also shown by [22] for the reference rivers in two European ecoregions. However, the values of the MTR system for the Solak station show some risk of eutrophication, which comes from the necessity of adapting the gradation between the ecological status classes in accordance with the local environmental conditions. This can be achieved via expanding studies to other available sites in the Caucasus and using general recommendations developed in the framework of the EU WFD [21]. This should be followed by the establishment of Ecological Quality Ratios (EQRs) for macrophyte-based assessment systems [77].