Twenty-four water reservoirs of the eastern region of Cuba, located in the provinces of Santiago de Cuba (10), Holguín (5), Granma (5), and Guantánamo (4) were sampled during September, 2015 to April, 2017 (figure 1). These reservoirs belongs to different basin and microbasin systems; having different capacities (ranged between 2, 38 to 330, 00 x 10⁶ m³) being the main uses: crop irrigation and water supply for domestic and industrials purposes, benefiting around 3 625 800 inhabitants in the eastern region of Cuba.

Fig. 1

TP (mgL^-1), Chla concentrations (µgL^-1) and SDD (m) were measured. Data collection was limited to surface water samples (first 50 cm of water column). Four location were established in each reservoir. Three measurements were done in each location.

Chlorophyll a was measured in situ using an Aquafluor^TM handheld fluorometer (Turner Designs, USA) previously calibrated with chlorophyll a standard (Sigma) (Chla_f). Adjustment to express results of chlorophyll a in vivo as µgL^-1 was made determining a correction factor, after the development of a calibration curve, which was prepared using serial dilutions from a standard Chla solution in methanol 90% (10 to 1000 µgL^-1).¹⁶ It was developed measuring the absorbance (A) at 675 nm, using a spectrophotometer UV-vis (Genesys, USA). The concentrations of the samples were determined considering the resulting curve equation.

To determine TP concentrations, samples were collected in triplicate in acid-prewashed leak-proof polyethylene sample bottles into the first 50 cm of water column; sample bottles have a double seal to ensure safe transportation; they were maintained at 4 ºC until analyses with a maximum storage time of 2 days. TP was determined spectrophotometrically with a Genesys 10 UV-vis spectrophotometer, after persulfate oxidation to phosphate ¹⁷, using the ascorbic acid method.¹⁸

To calculate the TSI values, the Carlson and Cunha et al. indexes were compared in order to select one criterion.^6-11

Carlson index considers the TSI average (TSI_a) calculated from measurements of SDD, Chla, and TP (n=96). The equations to calculate each TSI are showing below (Equation 1-3):

Cunha index considers the TSI average (TSI_tsr) calculated from measurements of chlorophyll a (Chla), and total phosphorus (TP) (n=96). The equations to calculate each TSI are showing below (Equation 4-5):

Classifications of the water reservoirs according TSI values were done according to both criteria, Carlson and Cunha et al. ^6,11,19,20

Samples were collected into the first 50 cm of the water column, using borosilicate containers of 250 mL, none preserved, but transported to the laboratory in complete darkness and cooling, proceeding immediately to their counting and cyanobacteria identification.⁷

Counting of cells (x 10⁴ cell.mL^-1) was made in fresh samples using a haemocytometer Neubauer. A differentiated counting of cyanobacteria was developed.

Taxonomic identification of cyanobacteria was fulfilled with different criteria till species when it was possible.^21-28 The list of species will be presented, highlighting the toxic species.

During the sampling some qualitative analyses were taking into account, based on determinations of some key indicators or attributes of the reservoirs like water discoloration, presence of algal bloom and algal scum; odor changes, macrophytes presence, and toxic cyanobacteria species occurrence.^19,29

Statistical analyses (mean, standard deviation, and relative error), all considered P˂0,05. These were conducted using R studio (v 1.1.456) Sigmaplot (v 11.0.0.77) and Origin Pro 9. To analyse the quantitative similarity between samples, attending to TSI average, a parametric analysis between samples were done with resemblance measure using software Primer 6, considering a S18 Kulczynski, creating a lower triangular resemblance matrix previous to the cluster analysis.

Pair samples test with two tailed of significance was conducted to compare TSI obtained by both methodologies. It was included a Pearson Correlation analysis.

According to the calibration curve was possible to adjust the results of chlorophyll a in vivo using the resulting curve equation (y= 0, 001 3 x + 0,001 7; R²=0,996 1) where y is A₆₇₅ and x is the concentration (µgL^-1) (figure 2). The correction factor was calculated for each sample from the ratio Chla _f: Chla DO ₆₇₅, resulting an average of 6,902 5 ± 0,000 1. The limits values of Chla (µgL^-1) were analysed, ranged between 19, 28 to 127, 43 µgL^-1.

Fig. 2

The results of TSI values of the twenty-four reservoirs studied are presented in table 1, as well as the status according to different classification criteria.^6,11,19,20 Trophic State Index (TSI) is frequently used as biomass-related trophic state, and it is generally easy to understand, both in theory and practice, being useful for monitoring and management purposes, even punctually as a starting point for analysing data on water quality in the tropics/subtropics since this issue is of paramount importance worldwide.⁶ The TSI average (TSIa) is a better criterion to classify and characterize the trophic state of lakes and reservoirs than the use of separated index.²⁰

During this study was confirmed the eutrophic status of all the reservoirs, which classified in different levels of eutrophication according to Cunha and Carlson criteria.^6,11,19,20

Los Plátanos reservoir with TSI_a = 55,36 ± 1,02, classified into the level 1 according to Carlson; but according to Cunha et al. (TSI_tsr =55,78 ± 0,02) the reservoir is in the limit of mesotrophy and eutrophyc. In this reservoirs highlight low levels of Chla. Los Plátanos is a small water reservoir, located near to El Salvador community (downstream), at north of Guantanamo city. The reservoir is near to a mountain system with forest coverage.

The highest values of both TSIs corresponds to Camazán (77,02 ± 0,39 and 62,91 ± 2,33) and Chalóns (74,48 ± 2,38 and 62,95 ± 1,12). Camazán reservoir receives direct run-off and contribution from agricultural areas, and the river received the influence of human pressure because of the settlements and Chalóns reservoir, (north of Santiago de Cuba city) receive the direct impact of Boniato town, agricultural areas, and cattle raising, impacting the water quality (table 1).

Even when this study evaluates TSI average calculated according to Carlson and Cunha, other authors use TSI calculated only with Chla with similar results. In this study, the TSI average was a better criterion to evaluate TSI in water reservoirs.²⁰ These results need to be complemented with qualitative analysis, it was relevant to analyse the Mícara reservoir.

Table 1

According to Carlson, Mícara is considered into the eutrophic level 2, establishing as main attributes, the dominance of blue-green algae, algal scum probable, extensive macrophyte problems.^11,19 Nevertheless in Mícara was detected algae bloom accompanied by water discoloration. It means that this reservoir have features of level 3 following Carlson criteria, but not correspond with attributes of eutrophic level 2.

All these analyses showed how is difficult to select one or other criteria to establish the eutrophic levels. Comparing media values of TSI obtained following Carlson and Cunha et al. equations ^6,11,19,20, these are not significantly different (P<0,05) (59,80 ± 4,26 and 60,07 ± 1,44, respectively), but Carlson is the most used index. It was the main reason to analyse the data considering a cluster analysis of TSI_a of Carlson classification.

The eutrophic conditions of water reservoirs are related to catchment uses, catchment populations linked with anthropogenic pressures, mainly because water polluted discharge, in those ecosystems resulting in inadequate management: A problem that needs to be solved.

Another important pressure is the aquaculture development, observed in all studied water reservoirs, due to fertilization practices which provide nutrients that condition the eutrophic status.

Considering the possibility to adjust the classification attending to the specific environmental scenario and features of the reservoirs, a cluster analysis was conducted in order to group the reservoirs based on TSIa similarities, then the coincidence with TSIs ranges used in the Carlson classification was evaluated (figure 3). It allows considering at least four eutrophic levels but a re-assignation of TSIa values was necessary, as well as new attributes specification.

Fig. 3

As a result of this research, the Carlson criteria were modified, mainly in the eutrophic ranges (≥50). An extra level (L5) was proposed taking into account TSI values higher than 78 (the maximum TSI value detected during this study), was potentially possible.

To establish the attributes, qualitative and quantitative indicators were taking into account to analyse the reservoirs attributes like phytoplankton and cyanobacteria counting, and some qualitative analyses like algal bloom and algal scum presence; as well as odour changes, macrophytes presence, and toxic cyanobacteria species occurrence were useful (table 2).

Table 2

Cyanobacteria blooms were detected in 37 % of total reservoirs ⁹, these were distributed in level 2 (1 reservoir), level 3 (6 reservoirs) and level 4 (2) of eutrophy. In all of them toxic bloom-forming species were involved.

A bloom of toxic species: Cylindrospermopsis raciborskii and M. flos-aquae was identified in Hatillo (level 2).^30,31 In this level 2, the bloom frequency was 20 %.

In the 16 reservoirs classified into the level 3, blooms were detected in six reservoirs (38 %), involving different toxic species: Microcystis aeruginosa (Parada and La Yaya), C. curvispora and Oscillatoria sp. (Baraguá), M. aeruginosa, M. flos-aquae, M. panniformis and M. wesenbergii (Mícara), Planktolyngbya limnetica and Oscillatoria sp. (Bio) and Synechococcus sp. (Guisa).

The bloom frequency in water reservoirs classified as level 4 was 100 %, being Synechocystis aquatilis (Chalóns) and C. raciborskii (Camazán) the species involved. Specifically, in these reservoirs, unpleasant odours, water discoloration, and dead fish were detected.

Scums were present in 46 % (11) of all the reservoirs, being the level 3 and 4 of eutrophication the most affected, mainly related to the bloom occurrence; highlighting the lower presence of macrophytes on these reservoirs (table 2).

According to these results, TSI can be easily used to predict short-term trophic dynamics of the studied reservoirs, even when not provide data concerning phytoplankton composition, but it is necessary to take into account. There are some researches that link the trophic status of reservoirs with cyanobacteria blooms, which support these results.³²

Phytoplankton cell density (table 2) ranged from 2,00 ± 0,20 x 10⁴ cell.mL^-1 (Gota Blanca) to 26,0 ± 0,39 x 10⁴ cell.mL^-1 (Parada). The average values have an increasing tendency in relation to trophic level, being 6,9 ± 0,52 x 10⁴ cell.mL^-1 in level 2, 8,98 ± 0,67 x 10⁴ cell.mL^-1 in level 3 and 9,37 ± 0,26 x 10⁴ cell.mL^-1 in level 4. The reservoirs with the higher density of phytoplankton cells were Parada, Mícara, Hatillo, la Yaya, Guisa and Chalóns (table 2). The phytoplankton concentration was in the range of previous reports in studies conducted in tropical reservoirs, with high levels of cyanobacteria and cyanotoxins ^33,34

Cyanobacterial concentrations (table 2) ranged from 0,50 ± 0,01 x 10⁴ cell.mL^-1 to 9,84 ± 0,02 x 10⁴ cell.mL^-1, being the highest average concentrations those corresponding to level 4 (6 ± 0,38 x 10⁴ cells.mL^-1), which was significant different with the rest of levels. Concentrations of cyanobacteria range from 1,22 ± 0,02 x 10⁴ cells.mL^-1, even though, there are no significant variation between concentration of cyanobacteria between levels 2 and 3 (2,64 ± 0,33 and 3,46 ± 0,45 x 10⁴ cells.mL^-1, respectively), there is a significant differences in relation with level 1. There is a correlation between eutrophication level and the concentration of cyanobacteria.

The results suggest that the risk because of cyanobacteria occurrence is closely related to the eutrophic level, being imminent in reservoirs classified in level 4 (table 2). High loads of nutrients and other factors can condition the eutrophication and collaterally the growth of potentially toxic cyanobacteria blooms is detected. The mechanisms underlying cyanobacterial growth and bloom-forming need further research. A general analysis allows to propose a new trophic state classification, with TSI values and attributes modified according to the results of the cluster analysis of the study reservoirs, it has consistent differences in relations with Carlson classification (table 3).

Table 3

TSI values	Trophic status	Attributes
<58	Eutrophic (level 1)	Lower boundary of classical eutrophy: decreased transparency, abundance of cyanobacteria, low level of total phosphorus, good for fisheries, acceptable to water supply.
58-64	Eutrophic (level 2)	Decreased transparency, cyanobacteria could be dominant, algal scum or algal bloom probable, extensive macrophyte problems, low risk of toxic cyanobacteria. Acceptable for fisheries and water supply.
>64-73	Eutrophic (level 3)	Dominance of cyanobacteria, algal scum probable, heavy algal bloom probable. Extensive macrophyte problems. Medium to high risk of toxic cyanobacteria. Limited quality for fisheries and water supply.
>73-78	Eutrophic (level 4)	Heavy algal blooms, algal scum, fish kills, few macrophytes. High risk of toxic cyanobacteria. Restrictions for fisheries and water supply.
>78	Eutrophic (level 5)	Higher boundary of eutrophy. Persistent bloom of toxic cyanobacteria, fish kills, bad odours, bacterial growth. Fisheries and water supply prohibited.

In all the studied reservoirs were identified 38 species of cyanobacteria. The reservoirs with higher cyanobacteria diversity were Baraguá (20), Bio (16) and Camazán (16) (table 2).

Toxic species represent the 61 % of total cyanobacteria, highlighting the occurrence of toxic and bloom-forming species, which are usually involved in episodes of animal and human toxicity worldwide (Microcystis sp., Cylindrospermopsis sp., Planktolyngbya sp., Oscillatoria sp., Synechococcus sp. and Synechocystis sp.).^30,35-37

The most frequent species were Synechocystis aquatilis and Aphanocapsa sp. with a spacial frequency of 62, 5 %, followed by Planktothrix sp., Merismopedia punctata (54 %) and Romeria simplex (50 %). Microcystis genus was the better represented with six species which are present in 70 % of the reservoirs

Concerning toxic species diversity was higher in Baragua reservoir (13 species), which represents 65 % of the total cyanobacteria species in this reservoir; followed by Bio, Cautillo, and Camazán which have 10 toxic cyanobacteria species, representing a 62, 67 and 62 %, respectively. It is an important rick because of the potential productions of cyanotoxins (table 4).

Table 4

Eutrophic Level	1	2				3
Reservoir	Los plátanos	Gota Blanca	Hatillo	Joturo	Paso Malo	Clotilde	Céspedes	Gilbert	Charco Mono	Parada	Mícara	Baraguá
Cyanobacteria species
Aphanizomenon flos-aquae						x						x
Aphanizomenon gracile	x						x					x
Aphanocapsa sp.	x	x		x	x		x	x	x		x
Cylindrospermopsis curvispora			x									x*
Cylindrospermopsis raciborskii			x*									x
Gomphosphaeria sp.		x	x						x
Leptolyngbya sp.												x
Limnothrix sp.												x
Merismopedia punctate	x			x	x	x			x		x	x
Microcystis aeruginosa										x*	x*
Microcystis comperei				x			x
Microcystis flos-aquae			x*								x*
Microcystis panniformis			x				x				x*	x
Microcystis viridis				x				x	x	x		x
Microcystis wesenbergii								x			x*
Oscillatoria sp.		x					x	x				x*
Planktolyngbya limnetica						x						x
Planktolyngbya sp.	x		x				x	x
Planktothrix sp.	x			x				x	x		x
Pseudanabaena sp.												x
Synechococcus sp.								x
Synechocystis aquatilis	x					x		x	x	x	x	x
Woronichinia sp.									x
Aphanizomenon flos-aquae	x					x		x	x	x		x
Aphanizomenon gracile								x				x
Aphanocapsa sp.		x				x	x	x	x	x	x
Cylindrospermopsis curvispora								x		x		x
Cylindrospermopsis raciborskii	x											x
Gomphosphaeria sp.		x
Leptolyngbya sp.	x											x
Limnothrix sp.	x		x									x
Merismopedia punctate	x				x		x		x	x		x
Microcystis aeruginosa			x		x	x*	x				x
Microcystis comperei			x								x
Microcystis flos-aquae					x	x		x	x	x
Microcystis panniformis										x		x
Microcystis viridis		x	x	x	x
Microcystis wesenbergii		x		x	x						x
Oscillatoria sp.	x*	x			x
Planktolyngbya limnetica	x*				x		x
Planktolyngbya sp.								x
Planktothrix sp.	x	x	x	x			x		x	x	x
Pseudanabaena sp.	x	x		x	x							x
Synechococcus sp.			x*	x	x						x
Synechocystis aquatilis	x	x		x	x	x		x			x*	x
Woronichinia sp.

Legend:*= Bloom