Cyanobacterial blooms are becoming more frequent and widespread in freshwater reservoirs worldwide, mainly due to anthropogenic activities and to climate changes [1
]. Some toxic cyanobacterial species are invading and expanding to new geographic regions [1
]. Considering the ability of cyanobacteria to produce several types of cyanotoxins, the monitoring of freshwater reservoirs is crucial to prevent acute and chronic health effects [1
]. This is particularly important in those reservoirs used as a source of drinking water and for recreational activities, where people are exposed to cyanotoxins by oral, dermal and inhalation routes [1
Portuguese surface freshwater resources are largely used for energy production, irrigation, recreational activities and as a source of drinking water for human consumption [4
]. An important proportion of those resources are river-derived artificial reservoirs, such as water dams. These artificial barriers alter the natural river flow and create lentic water bodies that are vulnerable to eutrophication [4
], which favors cyanobacteria proliferation [5
]. In fact, cyanobacterial blooming in Portuguese surface freshwaters is common and often associated with cyanotoxin production [6
]. Thus, the monitoring of cyanobacteria and cyanotoxins is a major issue for public health.
The Portuguese legislation for drinking water, Decreto-Lei nº306/2007 [14
], transposed from the European Drinking Water Directive [15
], established the regulatory level of 1 µg/L for total microcystins in treated water. This parameter is determined when eutrophication of water is suspected and when the number of potentially toxic cyanobacteria exceeds 2000 cells/mL. Conversely, the Portuguese legislation concerning the quality of bathing water [16
], that was transposed from the European Bathing Water Directive [18
], did not include any guideline for cyanobacterial cells or microcystin concentrations. It only recommends that when the bathing water profile indicates a potential for cyanobacterial proliferation, appropriate monitoring should be carried out to enable timely identification of health risks. When cyanobacterial proliferation is detected visually, it is the responsibility of the local health delegate to evaluate the health risk. If any risk has been identified or presumed, the health and environmental authorities should implement the adequate management measures to prevent exposure, including information to the public.
According to national specificities and circumstances, some European countries complemented the European Bathing Water Directive [18
] and implemented their own guidance or regulations, based on cyanobacterial cell numbers, biovolumes, pigments and/or cyanotoxin concentrations [19
The lack of a national database compiling the results of cyanobacteria and cyanotoxin monitoring hinders the characterization of Portuguese inland waters concerning toxic cyanobacterial blooms and the associated risks for public health and, in consequence, the implementation of guidelines.
Furthermore, we must take into account that the use of those reservoirs is increasing, not only for the production of drinking water, but also for ludic activities, such as water sports, fishing, sailing, swimming and bathing. In addition, the vicinities of those reservoirs host camping, fairs, national and international summer music festivals, natural parks and hunting reserves. Thus, the potential risks associated with these reservoirs must be evaluated and managed in order to protect the health of their users.
In this work, we compiled the results obtained during 14 years of monitoring of cyanobacteria and microcystins in eight freshwater reservoirs located in the centre of Portugal used for bathing and recreational activities. Based on these data, we profiled cyanobacterial blooms and determined the risk levels of toxic cyanobacteria occurrence and putative human exposure scenarios in these reservoirs.
The study of the occurrence of cyanobacteria in Portuguese freshwater reservoirs dates back to the 1930s, but it was only in the 1990s that research on the toxicity and distribution of toxic species gained special emphasis, with the works of Vasconcelos and collaborators, particularly in the north of the country (revised in [21
]). In the late 90s, monitoring programs of cyanobacteria and cyanotoxins in surface freshwater reservoirs were implemented by the Portuguese health authorities [7
]. However, if the quality of drinking water is, to some extent, safeguarded by Portuguese legislation [14
], the national legislation concerning bathing waters do not include yet parametric values for cyanobacteria and microcystins.
With this work, we intended to evaluate the trend of cyanobacterial blooms and microcystins in eight freshwater reservoirs used for recreation and determine the risk levels and putative human exposure scenarios associated with those blooms.
The results show that the studied freshwater reservoirs have distinct profiles of cyanobacteria proliferation and microcystins occurrence and, consequently, different risk scenarios concerning the potential human exposure to cyanotoxins. Three reservoirs (A1, A2 and A3), did not present a potential risk of human exposure to microcystins. Although a low risk based on cyanobacteria density was observed in more than 50% of the samples and despite potentially toxic cyanobacteria occurred, mainly Microcystis aeruginosa
, microcystins were never detected in water samples. These reservoirs are classified as mesotrophic (A1 and A3) and oligotrophic (A2), which might explain the low cyanobacteria density along all the monitoring period (from 9 to 14 years). Besides, non-toxic Microcystis aeruginosa
strains might have been predominant within the cyanobacterial communities in these reservoirs. In three reservoirs (A4, A5 and A6), the occurrence of microcystins represented a low risk of potential acute effects because toxin levels were always below 4 μg/L and the water volumes required to produce such effects in adults were quite improbable. However an accidental swallowing of 100 mL of water in reservoirs A5 and A6 could already represent a health risk in children weighing up to 10 Kg. In addition, the risks resulting from repeated exposures should not be disregarded, particularly in the eutrophic reservoir A6, where the risk level based on cyanobacteria density was high and where blooms were quite persistent. Potential toxic genus (Aphanizomenon
) were dominant and, inclusively, two strains of M. aeruginosa
isolated from reservoir A4 in October 2003 and from reservoir A5 in September 2002 were previously reported as producers of several microcystin variants (MCLR, MCRR and MCYR) [23
]. Thus, the occurrence of microcystins at those time points was, at least, attributed to the species M. aeruginosa
. Although no cyanobacterial strains were isolated from reservoir A6 during the monitoring period, this reservoir has a previous history concerning the occurrence of toxic strains of M. aeruginosa
, producers of several variants of microcystins [23
]. In two reservoirs (A7 and A8), the high risk level of 20 µg MCLR/L established by WHO [26
] was exceeded in 7% of the samples, reaching a maximum of 506 µg/L in reservoir A7 and 389 µg/L in reservoir A8. However, the reality of these two reservoirs was quite different. The highest concentrations of microcystins in the reservoir A7 were detected only in the period of August–October 2010, remaining at residual levels (bellow 1 µg/L) in the other periods. In reservoir A8, microcystins peaked in distinct years and persisted along the monitoring period at residual levels or between 1 µg/L and 20 µg/L. In addition, the risk level based on cyanobacterial density was in general low at reservoir A7, whereas a high-risk level was determined in 51% of the samples from reservoir A8. In both reservoirs, realistic microcystins exposure scenarios were identified, more often in reservoir A8. Indeed, in some toxic bloom occurrences, only very small volumes of ingested water (1–5 mL) would be required to induce harmful health effects. Despite the great diversity of cyanobacteria in these reservoirs, the production of microcystins has only been confirmed to date in M. aeruginosa
strains isolated from reservoir A8 [23
Overall, we found a great diversity of species, being the most prevalent, among the potential toxic genera, Microcystis
, Dolichospermum, Aphanizomenon
. Previously, other authors also reported the prevalence and dominance of Microcystis
spp. in freshwaters from the north [27
] and south [10
] of Portugal. It should be noted, however, that blooms of Dolichospermum
(formerly named as Anabaena
] and Planktothrix
] have also been reported in Portuguese freshwaters. Other studies showed that the MCLR variant is the most common, with a proportion of 44 to 100% in relation to the total microcystins content [6
]. It was also concluded that MCRR, MCYR, MCAR and [D-Asp3] MCLR also occurs, but in a lower proportion suggesting that the methylated variants of microcystins are the most common in Portugal [6
]. Additionally, other cyanotoxins have also been detected in blooms or in cultured strains isolated from natural samples, such as saxitoxins [8
] and anatoxin [36
]. Besides, if it was once considered that some cyanobacteria species were confined to certain geographical locations, namely tropical regions, nowadays those species and their toxic compounds are emerging and spreading into other regions. It is the case, for example, of Cylindrospermopsis raciborskii
] and the amino acid β
-alanine (BMAA) [38
] that have been detected in reservoirs and/or estuaries in several European countries, including Portugal. In reservoirs A6 and A8, C. raciborskii
also occurred sporadically, but its toxicity was not evaluated. This points out that the risk levels in the studied reservoirs may be underestimated considering that cyanobacterial species that are potentially producers of other toxins also occur frequently and at high cell densities.
We should also emphasize that, besides the acute cases of illness in humans exposed to microcystins during recreational activities [39
], continuous exposure to water contaminated with low concentrations of microcystins may also pose a risk of chronic health effects [1
]. Thus, the health risks resulting from human and animal exposure to cyanobacterial toxins should not be disregarded in five of the eight evaluated recreational freshwater reservoirs (A4 to A8). Indeed, although recreational exposure is intermittent by nature, the bathing season in this region may start in early spring and last until early autumn. Additionally, other activities besides bathing, such as water sports, occur along the year. Furthermore, exposure routes other than ingestion should be considered, such as aerosol exposure to low concentrations of microcystins (2 to 5 μg/L) during recreational activities [42
On the other hand, it has been reported that cyanotoxins may accumulate in horticultural crops [44
] that might have negative consequences to the physiology of the plants as well as to the health of the consumers. Thus, potential economic loses and human/animal health risks should also be considered in Portuguese surface freshwater reservoirs, since many are also used for irrigation, such as reservoirs A5, A7 and A8.
The available epidemiologic data concerning human exposure to cyanotoxins in Portuguese freshwaters is restricted to two situations. The first occurred in 1993 in a dialysis unit in Évora Hospital (Alentejo, south Portugal), where twenty dialysis patients died [45
]. This was an episode very similar to the well-known Caruaru incident in Brazil. The water that supplied the city was heavily contaminated with Microcystis
, but it also contained high levels of aluminum [45
]. Unfortunately, cyanotoxins analyses were not performed in the victims or in the water, and the correlation with cyanobacteria remained uncovered. However, an epidemiological study of ecological type showed a correlation between the occurrence of cyanobacteria and the change in liver enzymes in the population supplied by the contaminated water reservoir. This was the first time that the occurrence of cyanobacteria was considered as a risk factor to human health in Portugal [45
]. A more recent study from Bellém et al. [13
] alerted for the potential public health concern in six freshwater reservoirs from Alentejo (Alvito, Boavista, Enxoé, Roxo, Monte Novo, Vigia), considering the dominance of potentially toxic cyanobacteria from spring to late autumn, with some recorded cell density values corresponding to medium/high risk factor, according to the guidelines from WHO [26
]. Unfortunately, this paper does not mention the levels of cyanotoxins in those reservoirs. Nevertheless, the same author developed a historical prospective study (2000–2010), in which two populations from Alentejo [one supplied by freshwater reservoirs contaminated with cyanobacterial blooms (exposed population) and one supplied by other water sources (unexposed population)] were compared relatively to the laboratory indicators of hepatic pathology. According to the author, the exposure to cyanotoxins had an impact on liver disease, attested by the high values of transaminases and the higher incidence of hepatocarcinoma in the exposed population [47
]. There is no national data on adverse health effects resulting from the exposure to cyanotoxins during recreational activities, but those previous reports point out the potential risk associated with surface freshwaters reservoirs from south Portugal. These regions have a Mediterranean weather, with mild winters and hot and dry summers, being especially vulnerable to climate changes, and severe or extreme drought in particular, which may favor eutrophication and cyanobacteria blooming [5
It was not our goal with the present paper to draw conclusions about the causes of the blooms in the studied reservoirs. Indeed, the monitoring period was not the same among them and, consequently, the seasonality of blooms in the reservoirs can hardly be compared. Nevertheless, we observed that, in a general way, the periods of 2000–2002 and 2007–2009 corresponded to the highest bloom frequencies. It is interesting to note that these dates were preceded by periods of extreme droughts, namely in 1998–1999 and 2004–2006 [51
]; the latest, inclusively, was the most extreme drought in terms of territorial extension of the last 65 years. We can hypothesize that these climatic extremes may have provided the conjugation of biotic and/or abiotic factors that, in turn, triggered the increase of blooms in subsequent years.
On the other hand, although the blooms were detected mainly in spring and summer seasons, the reservoirs A4 to A8 also presented blooms in autumn and winter. These blooms occurred when the average air temperature was above normal for that period (e.g., January 2003, November/December 2007, November 2009) or in rainy months that were preceded by heat waves and/or dry summers (e.g., October 2003, October 2011) [51
]. This is similar to what has been previously described in Mediterranean areas, such as Greece, where the blooms may last until December or even persist throughout the year [49
]. Most of those blooms were composed by filamentous species, namely heterocysts and akinetes forming species, such as Aphanizomenon
spp. and even Cylindrospermopsis raciborskii
, probably due to their highly successful adaptive strategies.
Many countries that regulate cyanotoxins in drinking water use a parametric value based on the WHO Guideline Value for MCLR of 1 μg/L [20
]. However, and according to regional specificities, some countries also adopted guidance/alert/maximum levels for cylindrospermopsin (Australia, New Zealand, Brazil), saxitoxin (Australia, New Zealand, Brazil) and anatoxin-a (Canada, New Zealand) [19
]. In what concerns recreational waters, most countries use guidance values based on cyanobacterial biomass (cell density, chlorophyll-a, biovolume) reflecting, indirectly, potential hazardous microcystin concentrations [19
]. The risk of health problems due to the presence of other cyanotoxins in bathing waters is recognized, but the adoption of guidance values for those cyanotoxins in water bodies used for recreation is still an issue under discussion [19