The Valsequillo reservoir, located near the city of Puebla, Mexico, is a hard-water eutrophic subtropical system, with minimum temperatures in winter (November–December), and marked dry-rain seasons with fluctuating depth. The reservoir has been infested with water hyacinth for over three decades. A management program involving the use of triturating machines was applied from December 1996 to February 1997. After trituration, remains were allowed to settle to the bottom. The purpose of this study was to monitor the changes in the water quality and the biological communities before and after physical control of weeds. A monthly sampling of surface water was performed at four stations one year before the treatment. After trituration, one year sampling was also carried out. Variables measured included temperature, Secchi disk transparency, depth, pH, dissolved oxygen, oxygen saturation, hardness, nitrate content, nitrite content, ammonia, orthophosphates, and numerical abundance of phytoplankton, zooplankton, and nekton.
Weed control affected changes in all variables measured, as a result of residual decomposition of triturated matter. Secchi transparency and oxygen levels decreased and pH became slightly more alkaline. More important changes occurred for nutrients. Orthophosphate concentration increased, for nitrate and nitrite, increase was about 320% and 450% respectively. Ammonia reached lethal values for at least four months after trituration. As a result, phytoplankton decreased initially, and when it flourished again, the Bacillariophyta were replaced by Cyanophyta. Euglenophyta were important in both years. Of zooplankton, calanoids decreased, but cyclopoids and cladocerans maintained similar numbers, although the latter group changed in composition in that Ceriodaphnia was replaced by Moina. Fish disappeared from the system after weed trituration. In the second year a small recovery of water quality occurred, but water hyacinth also started to develop again. At present, Valsequillo is again covered by water hyacinth.
Weed infestation is common in many Mexican reservoirs, affecting over 40 000 ha of dams, lakes, canals, and drains (Gutiérrez et al., 1996). Among the main problems caused by this phenomenon are the obstruction of irrigation channels, favoring development of harmful fauna of mosquitoes, obstructing navigation, and so on. In spite of all these problems, perhaps the major impact of water hyacinth is through evapotranspiration. Data from National Commission of Water (CNA) in Mexico published in a local newspaper (El Sol de Puebla, March 22, 1998), indicated a negative economic impact of about US $1.6 million in 1997 by loss of water just in Valsequillo dam.
In Mexico, removal of freshwater weeds is usually by means of herbicides, through mechanical trituration, or by biological control, with emphasis on the first two methods (Gutiérrez et al., 1994). Most commonly, there are no monitoring programs to evaluate the post-trituration effects on the aquatic environment. Gutiérrez et al. (1996) proposed a water quality monitoring program prior and subsequent to control actions. According to their data, in three Mexican dams no noticeable effects were found in the benthos and plankton communities after weed control measures were taken.
Lugo et al. (1998) studied changes after use of diquat and 2.4-D amine in a reservoir of Mexico City to eliminate the water hyacinth Eichhornia crassipes. The main effect was a decline of all planktonic communities and dissolved oxygen levels, as a secondary effect of the decomposing residual biomass from the weeds.
Weed control methods have been applied in most cases with no evaluation of their effects on other communities. Hence, the aim of this work was to evaluate effects on the main physical and chemical variables, the plankton, and the fish communities before and after this process.
The Manuel Avila Camacho reservoir (also known as Valsequillo) is an example of a suburban fresh water system located near a major city of Mexico. It is located 5 km south from Puebla, between 18° 53′ and 18° 57′ N, and 96° 06′ and 98° 15′ W at 2100 m above sea level (Figure 1). It was constructed for irrigation purposes in 1941–46 by the Mexican government. The initial water storage capacity was 405 × 106 m3. It receives, from the northern side, permanent discharge from the Atoyac river, and intermittent discharge from the Alseseca river, the latter restricted to the rainy season in summer. These two rivers transport most of the domestic, agricultural, and industrial waste waters of the region into the system with an increase of nutrient load as the most notable effect. As a result, the water hyacinth, Eichhornia crassipes, has became a major problem, covering almost 70% of the reservoir surface for about 30 years (Aldama, 2000) (Figure 1). The National Water Commission (Comisión Nacional del Agua, CNA) decided to remove the water hyacinth from Valsequillo by mechanical trituration and subsequent sedimentation of the resulting material in the bottom of the reservoir.
Methods applied to remove the weeds from Valsequillo reservoir
Weed control was performed with three triturating machines of the ‘Retador’ type, each with a capacity to destroy 9.8 metric tons per day of water hyacinth (Aldama, 2000). Machines were used on a daily basis from December 1996 to February 1997. No herbicides were used previous to this period or subsequently.
Four sampling stations were established along the main axis of the reservoir. One was near the water entrance, two in the middle part and one near the dam (Figure 1). Sampling of water was carried out monthly during two years from the surface layer of water (the first 30 cm). The first year was from June 1994 to April 1995, before water hyacinth removal, and the second from May 1997 to April 1998, after mechanical removal of weeds. No sampling occurred during November 1997. Sampling was performed between 9.00 and 11:00 AM.
Physical and chemical variables measured included water and air temperature with a Taylor thermometer; Secchi disk transparency; maximum depth; pH with a field Conductronic pHmeter; dissolved oxygen (Winkler method); oxygen saturation with a nomogram; total hardness with titration with ethylenedinitrilotetraacetic acid; alkalinity by the acid binding capacity by titration with hydrochloric acid; nitrate and nitrite with the method of sulphanilic acid; ammonia by the indophenol method, and orthophosphate by the vanamolybdic acid method (the latter four techniques described in the Aquamerck manual). A one way ANOVA was used to compare both years of study for each physical and chemical variable.
Each collection set included the following: phytoplankton in glass bottles of 250 ml, fixed with lugol. To count phytoplankton, all colonies were considered as a single individual. Zooplankton was obtained by filtering 100 l of water through a net of 60 μ m mesh size and preserved with 4% buffered formaldehyde; fish were captured with a seine net of 1 × 10 m and 2 mm mesh, and preserved with 10% formaldehyde.
Zooplankton was identified with the keys of Pennak (1989). Phytoplankton was identified with the aid of Prescott (1973), Tiffany and Briton (1971) and Ortega (1984). Fish were identified with the aid of Miller (1974) keys.
Results and discussion
From the ANOVA test, no significant differences were found among temperature, Secchi disk, pH, dissolved and saturation of oxygen, hardness and ammonia. Significant differences were found for depth, alkalinity, orthophosphate, nitrate and nitrite (see Table 1).
Water temperature followed a similar pattern as air temperature. The system belongs in a subtropical category, with a minimum of 18°C in November–December (Hutchinson, 1969). In the 1997 to 1998 cycle, the minimum was in February (13.9°C), as a result of a colder winter (Figure 2A). However, there were no significant thermal differences between the two cyles studied. Similar thermal conditions were found by Lugo et al. (1998) in the Guadalupe reservoir.
Measurements of pH (Figure 2B) showed a tendency from neutrality to acidic readings. After hyacinth removal values were more variable but no significant differences were found in both periods of sampling (see Figure 3J). Lugo et al. (1998) found a slight increase of pH after water hyacinth removal, probably due to algal blooms.
Water depth values were typical for this kind of system, with maximum depth after the rainy (summer) and minimum values at the end of dry season (autumn–winter) (Umaña and Collado, 1990), and with the discharge of water from the reservoir for irrigation (Figure 2C). Generally, in the second year the depth increased significantly, probably due to the diminishing of evapotranspiration after removal of water hyacinth, because use of the water from the reservoir is more or less constant. Secchi transparency decreased after hyacinth removal from an average of 1.4± 0.8 m to 1.3± 0.7 m, but this difference was found to be not significant with ANOVA analysis. In both years, maximum Secchi transparencies were detected during the dry season, from February to May; and minimal ones occurred after the rainy season, from October to January (Figure 2D).
Dissolved oxygen and saturation followed a similar pattern, with lowest values in November, 1995 (0.4 mg l− 1). Although in 1997 to 1998 there was no record for this month, the general pattern was similar in both periods, with maximum readings in May and June (Figures 2E, F). Total average of dissolved oxygen was 3.7 mg l− 1 for the first period and 2.44 mg l− 1 for the second one, but no significant differences were found between the two years. According to our data, the surface oxygen decrease was not so drastic as the change detected by Lugo et al. (1998) after weed removal.
The readings of pH and total hardness indicate the Valsequillo reservoir is a medium hard to hard water system, according to the criteria of Hütter (1988), and hard water according to Reid (1960). Changes in hardness were minor after the removal of water hyacinth (Figure 2B), and no significant differences were found.
Orthophosphate concentration increased markedly, mainly at the beginning of the second period of study. High fluctuations on this parameter are related to the instability of the system following crushing of water hyacinth. Maximum values after removal were around 15–25 μ g At l− 1, but reached near 30 μ g At l− 1, exceeding values from other systems with induced eutrophy, such as the Catemaco lake, in Veracruz (Tavera and Castillo, 2000).
Ammonia increased to lethal values (Figure 2J), from less than 3 mg l− 1 to near 60 mg l− 1 at several sites. Maximum average peak was in May 1997 with 40 mg l− 1 NH4+, a lethal concentration for most aquatic forms. The increase in NH4+ was due to the high rate of decomposition of the triturated hyacinth along the water column and in the sediment. Although statistical differences were not found in the comparison of the two years, in the beginning of the second year an important increase of dissolved NH4+ in May was noticeable, with a rapid recovery after July to similar values of the previous year. We consider that the high values reached in the second year affected all living forms, mainly fishes, since they disappeared from the system. This decomposition also affected other forms of nitrogen available, for example, nitrate (Figure 3A), which increased about 320% on average and nitrite, with a 450% increase (Figure 3B). In both cases significant differences were found between the two years.
The composition and abundance of some groups of phytoplankton were modified after mechanical removal of water hyacinth. Diatoms (Bacillariophyta) were the most abundant in the pre-treatment year (Figure 3I), together with small numbers of Cyanophyta and Euglenophyta. By the post-treatment year, diatoms disappeared from May to October, and they were replaced by high numbers of Euglenophyta (with a predominance of Euglena acus and Phacus sp.), and toxic Cyanophyta (dominated by Microcystis aeruginosa and Oscillatoria lutea) (Figure 3J). The maximum peak of both species was observed in May 1997 (Figure 3J). Maximum counts were observed at stations II and III, with 6292 and 5399 cells ml− 1, respectively. Phytoplankton composition and dominance suggest eutrophy, but the quality of the water was poorer after the trituration process. A slow recovery started after November of the second year, when diatoms reappeared, and the total number of phytoplankton cells dropped to less than 600 cells ml−1.
The zooplankton community also changed. During the first year, calanoids were present at almost all times; in March they reached up to 5030 ind 100 l− 1, followed by April, with 4648 ind 100 l− 1. After weed removal, their number decreased to a maximum of 1291 ind 100 l− 1, but during all remaining months their abundance never exceeded 170 ind 100 l− 1 (Figure 3C). Cyclopoids were dominated by two genera, Macrocyclops and Acanthocyclops. Their numbers decreased from a maximum of 16872 ind 100 l− 1 in the first year to 3377 ind 100 l− 1 in the second year. Thus, it appears that copepods were more resistant to the environmental change. Their fluctuations were similar during both cycles (Figure 3D), although numbers were lower in the second cycle. Cladocera presented a similar pattern before and after triturating (Figure 3E), but the composition of this group changed from a dominance of Daphnia and Ceriodaphnia to a dominance of the more resistant Moina, which represented only about 1% of the cladocerans before trituration (Figures 3G–H). Also, the numbers of Daphnia and Simocephalus were diminished. An inverse relationship was found by Lugo et al. (1998); the dominance of Moina shifted to a dominance of Daphnia after weed removal. Although Daphnia remained as an important member of the zooplankton, its abundance changed from a maximum of 130430 ind 100 l− 1 at station IV in April 1995 to a maximum of 14963 ind 100 l− 1 at station III in March 1998. Also, it is important to note that all cladocerans were almost absent from the system at the beginning (June and July) of the second year of sampling, but they recovered rapidly by August.
The most affected community was the fish. The three species found (Figure 4A–B) before crushing practically disappeared in the second cycle, probably affected by the effect of the increase in nutrients and ammonia. We were not able to find even a single fish after removal of water hyacinth, although it is well known that Cyprinus carpio and Poecilia sphenops are species highly resistant to environmental changes. Also, Heterandria jonessi, an endemic fish for the region according to Miller (1974), disappeared from the reservoir. Nowadays, fish have started to re-colonize the system (Mangas-Ramírez, pers. obs.)
The Valsequillo dam has a problem of excess nutrient input and an organic overload favoring both, algae with increased numbers of Cyanophyta, and the coverage of water hyacinth over nearly 70% of its surface. Crude waste waters from urban settlements along the two tributaries and in the vicinity of the dam were drained into the reservoir.
The main problem after weed removal was the increase of decomposing organic matter, with a resultant increase in nutrient concentration, a diminution of dissolved oxygen and the increase of toxicants such as ammonia. The effect of these changes was a depletion of all living forms in plankton and nekton, with a recovery in the next year, with the exception of fishes. Although alternative management has been studied to dispose the residuals of this weed, as organic inputs to soils or as livestock feed (e.g., Woomer et al., 2000) here the remains were allowed to settle in the bottom.
Our results disagree with the report by Gutiérrez et al. (1994), in which the authors reported no changes on plankton, benthos, or fish after hyacinth removal, but it should be pointed out that they used mostly chemical control rather than triturating. Although results by Lugo et al. (1998) indicated an important shift on planktonic communities after use of chemicals, and a rapid recovery, the high nutrient load of the system continued in the same way. In half a year after the treatment, the Guadalupe dam was again covered by water hyacinth (A. Lugo, CYMA project, Faculty of Superior Studies Iztacala, National Autonomous University of Mexico, pers. comm.).
The solution of the problem of water enrichment and subsequent proliferation of weeds should be an integral management goal of the basin, by identification of the main sources of pollution and development of water treatment strategies in view of the main sources of organic and chemical pollution. Instead of this, Mexican political authorities, under the pressure of the society for an immediate cleaning of the system, decided just to triturate the water hyacinth and leave it to settle in the bottom of the reservoir. After this failed management, problems with water quality increased, and resulted in a clearly negative effect on the aquatic communities dwelling in the system. The problem with water hyacinth has also continued, and three new triturating plans were performed from 1998 to 2000. Today, water hyacinth still covers the Valsequillo reservoir. Successful management of these aquatic problems requires consistent, long-term control efforts, coordinated and administered on a regional basis, as Charudattan (2001) proposed. Valsequillo is an example of a limited choice of control method, it is a recurrent problem without solution to the present. To successfully control this weed a more integrated approach using biological control combined with other methods (e.g., herbicidal) is needed, as suggested by Center et al. (1999). Also a water and basin management plan and a good long-term monitoring program should be considered.
A. Handal-Silva, E. Suárez-Morales, J. Alcocer and G. Vilaclara critically reviewed the original manuscript. Personnel of the limnology and marine sciences laboratory at Autonomous University of Puebla assisted us with field collections. Comisión Nacional del Agua (CNA) allowed us to use their data of ammonia for February–April, 1995. Fondo Mexicano para la Educación Superior (FOMES) assisted to this work through grant 98-2218. Dr. N.N. Smirnov (Russian Academy of Sciences) gave valuable advice to the English version.