HydroMet Buoy - 2019-2022 water quality at 1-m depth

Figure 1 - Time evolution of the YSI EXO2 sonde immersion depth (programmed at 1 m depth to take measurements every 30 min) during the monitoring period. On several occasions, the sonde remained near the bottom (dammed by drifting fishing nets; or for unknown reasons). It is necessary to be attentive to send the sonde to the surface and program it to start the profiles again. Since June 2020, it did not happen anymore. Care must be taken to interpret the data.

Figure 2 - Temporal evolution of water temperature at 1 m depth (except for the periods during which it was at the bottom; see previous figure). The temperature follows a perfectly parabolic curve, with a maximum of 17 ºC at the end of January, and a minimum of 11ºC in July. Note: dates are presented as month/day/year format.

Figure 3 - Temporal evolution of conductivity (in µS/cm). On average, it remains at 1,455 µS/cm, except between October and November when the YSI EXO2 sonde was trapped at the bottom, reaching 1,176-1,258 µS/cm (not visible on the graph). This suggests the presence of fresher water at the bottom.

Figure 4 - Temporal evolution of pH. From June to November, we had a decalibrated sensor (its drift is observed) that was replaced by a new one on 0/12/2019, since when the measurements have been reliable. The pH fluctuated between 8.75 and 9.21, with an average of 9.00.

Figure 5 - Time evolution of the oxidation-reduction potential (ORP) in milli-volts. The sensor was recalibrated in October. However, the data are reliable from December onwards when they fluctuate between 190 and 270 mV, being 241 mV on average. 

Figure 6 - Temporal evolution of the dissolved oxygen saturation percentage. The sensor was recalibrated in October. Since January, it fluctuated between 67 and 114%, with an average of 98%. This behavior represents a very good level of oxygenation, in equilibrium with the atmosphere. Cycles (oscillations) of 2-4 weeks are observed. 

Figure 7 - Temporal evolution of dissolved oxygen concentration. For the same reason, the data are reliable from October onwards. From October to December it fluctuates between 8.5 and 11.0 mg/L. From January on, it fluctuates between 7.5 and 11.0 mg/L, with periods of 2-4 weeks. From March onwards, it gradually increases up to 12.0 mg/L in June. Corresponds to a good level of oxygenation, given the altitude of the lake.

Figure 8 - Temporal evolution of turbidity. In the dry period (until November) it remains between 0.1-0.4 NTU. It increases in the rainy period (December-February). From January on, it reached 1.1 NTU, then gradually decreased with oscillations until reaching 0.1-0.2 NTU at the end of March. From June onwards, it increased slightly. The average is 0.28 NTU, with fluctuations between 0.08 and 4.02 NTU. So the water was very clear. It can be noted that the pattern of turbidity evolution follows the pattern of chlorophyll-a concentration (Fig. 9). In other words, turbidity results mainly from phytoplankton biomass (not from suspended mineral particles). 

Figure 9 - Temporal evolution of total chlorophyll-a concentration (in RFU). In the dry period, it remained between 0.2-1.0 RFU (according to the calibration with the FluoroProbe bbe probe presented in Annex 5 of the 5th Quarterly Report, it is necessary to multiply the NTU x 8 to obtain the µg/L), i.e. equivalent to 1.6-8.0 µg/L (oligotrophic to mesotrophic). In the rainy period in January-February (due to nutrient inputs from river wastewater, in addition to atmospheric inputs from aerosols) it increases between 0.5-2.8 RFU, i.e. equivalent to 4.0-22.4 µg/L (predominantly mesotrophic-eutrophic). During the annual cycle, chlorophyll-a fluctuated between 0.10 and 2.78 NTU (i.e. 0.80 and 22.24 µg/L), with an average of 0.69 NTU (i.e. 5.52 µg/L). In comparison, during 1979-1980, for the entire Minor Lake, chlorophyll-a concentrations did not exceed 0.5 µg/L in the dry period and 2.0 µg/L in the rainy period, with a maximum of 5.0 µg/L (Lazzaro, 1981).

Figure 10 - Temporal evolution of phycocyanin concentration (photosynthetic pigment of cyanobacteria). In this period it remained below 0.35 RFU, with a maximum of 1.72 RFU. According to the manufacturer's calibration YSI, 1 RFU = 1 µg/L phycocyanin. The average was 0.14 RFU (=0.14 µg/L) which is a very low concentration. The concentration was highest in January, and increased gradually from April onwards.

Figure 11 - Temporal evolution of the fluorescent dissolved organic matter (fDOM) concentration. During the dry period it varied between 0.8-1.4 RFU. It increased slightly during the rainy period from January 2020, up to 1.8 RFU. This increase in January and February resulted from organic matter inputs from the rivers, particularly from the Katari basin, via the Cohana and Sehuenca rivers joining to form the Cumana lagoon, which flows into Cumana Bay. As of March, it stabilized around 1.3 RFU. 

Figure 1 - Time evolution of the immersion depth of the YSI EXO2 sonde (programmed to stay at 1 m depth to take measurements every 30 min) during the two years of monitoring. On several occasions the sonde stayed close to the bottom (dammed by drifting fishing nets; by the depth gauge control; or by unknown causes; we fixed the problem in 2021 when it was much less frequent). So, care must be taken to consider these events in interpreting the data.

Abbreviations: E (Enero) = January; F (Febrero) = February; M (Marzo) = March; A (Abril) = April; M (Mayo) = May; 

Figure 2 - Temporal evolution of water temperature at 1 m depth (except for some periods during which it was at the bottom; see previous figure). The temperature followed a perfectly parabolic curve, with a maximum of 17.7 °C from January to March, and a minimum of 10.3 °C in June-July. The average for the period was 14.2 ºC and the median 14.4 ºC. During the 1st semester the temperature was slightly higher (≤ 1 ºC) in 2020 relative to 2021.

Figure 3 - Temporal evolution of conductivity (in µS/cm) at 1 m depth. On average it has remained at 1,455 µS/cm, except between October and November when the YSI EXO2 sonde was trapped at the bottom, reaching 1,176-1,258 µS/cm (not visible on the graph). This could suggest the presence of fresherwater at the bottom. From January to April, conductivity was higher in 2021 relative to 2020 during the same period. We do not know the cause of this phenomenon: would it be higher evaporation, lower rainfall and/or lower water level in 2021?

Figure 4 - Temporal evolution of pH at 1 m depth. From June to November 2019, we had a decalibrated sensor (its drift is observed) that was replaced by a new one on 10/12/2019, from when the measurements were reliable. The pH fluctuated between 8.75 and 9.21, with an average of 9.00 during 2020. However, from February 2021 onwards, the sampler became decalibrated again and it was not possible to calibrate it at all. So since February 2021 we have no pH data (nor ORP data for that matter). A new sensor was ordered; however, due to the shortage of components during the pandemic, it only just arrived on September 28, 2021. It will be installed in October.

Figure 5 - Time evolution of the oxidation-reduction potential (ORP) in milli-volts at 1 m depth. The sensor was recalibrated in October 2019. However, the data were only reliable from January 2021 when they fluctuated between 190 and 270 mV, being 241 mV on average. In 2021, the data has to be discarded. The sensor will be replaced in October 2021.

Figure 6 - Temporal evolution of the percentage of dissolved oxygen saturation at 1 m depth. The sensor was recalibrated in October 2019. As of January 2020, it fluctuated between 67 and 114%, with an average of 98%. This behavior represents a very good level of oxygenation, in equilibrium with the atmosphere. Cycles (oscillations) of 2-4 weeks are observed. Higher oxygenation of up to 20% more was observed during 2020 relative to 2021, except from October-December 2020 when oxygenation was lower by up to 20% to 2019 oxygenation. Higher oxygenation may also result from higher micro-algal biomass due to higher photosynthetic intensity.

Figure 7 - Temporal evolution of dissolved oxygen concentration at 1 m depth. Of course, it follows the same pattern as the % saturation in dissolved oxygen (%DO) in Fig. 6. For the same reason, the data are reliable as of October 2019. From October to December 2019 it fluctuated between 8.5 and 11.0 mg/L. From January 2020 it ranged between 7.5 and 11.0 mg/L, with periods of 2-4 weeks. From March 2020, it gradually increased to 12.0 mg/L in June. This corresponds to a good level of oxygenation, given the altitude of the lake. The concentration in dissolved oxygen was higher in 2020 by up to 2 mg/L relative to 2021, except from October to December when dissolved oxygen was lower than in 2019. As for %DO it may result from higher biomass/photosynthesis of micro-algae.

Figure 8 - Temporal evolution of turbidity at 1 m depth. In the dry period (until November) it remained between 0.1-0.6 NTU. It increased in the rainy period (December-February). From January 2020, it reached 1.1 NTU, then gradually decreased with oscillations until reaching 0.2-0.4 NTU at the end of April 2020. From June onwards, it increased slightly. The average is 0.34 NTU and the median is 0.32 NTU, with fluctuations between 0.08 and 4.02 NTU. In other words, the water was very clear. It can be noted that the pattern of turbidity evolution followed the pattern of chlorophyll-a concentration (Fig. 9). This means that the turbidity mainly reflects the phytoplankton biomass (but not suspended mineral particles).

Figure 9 - Temporal evolution of total chlorophyll-a concentration (in RFU) at 1 m depth. During the dry period, it remained between 0.2-1.0 RFU (according to the calibration with the FluoroProbe bbe probe presented in Annex 5 of the 5th Quarterly Report, one has to multiply NTU x 8 to obtain µg/L), i.e. equivalent to 1.6-8.0 µg/L (oligotrophic to mesotrophic). In the rainy period in January-February (due to nutrient inputs from river wastewater, in addition to atmospheric inputs from aerosols) it increased between 0.5-2.8 RFU, i.e. equivalent to 4.0-22.4 µg/L (predominantly mesotrophic-eutrophic). During the annual cycle, chlorophyll-a fluctuated between 0.1-3.0 NTU (i.e. 0.8-24.0 µg/L), averaging 0.34 RFU (i.e. 2.72 µg/L). In comparison, during 1979-1980, for all of Lake Minor, chlorophyll-a concentrations did not exceed 0.5 µg/L in the dry period and 2.0 µg/L in the rainy period, with a maximum of 5.0 µg/L (Lazzaro, 1981). This means that, compared to 1979-1980, the current Chl-a concentration is up to 16 times higher during the dry period, and up to 11 times higher during the rainy period, compared to 40 years ago. This demonstrates the significant increase in the biomass of micro-algae in response to the combination of the effects of climate change with several decades of massive inputs of nutrients from pollution from the urban area of El Alto through the Katari watershed.

Figure 10 - Temporal evolution of phycocyanin concentration (photosynthetic pigment of cyanobacteria) at 1 m depth. During this 2-year period, phycocyanin remained below 0.4 RFU, with a maximum of 2.9 RFU. According to the manufacturer's YSI calibration, 1 RFU = 1 µg/L of phycocyanin. The average was 0.78 RFU (=0.78 µg/L) which is a relatively low concentration. The concentration was slightly higher during the 2nd half of 2020 relative to 2019. It was slightly higher during the 1st half of 2021 relative to 2020, with a delayed peak at the beginning of March 2021 (0.4 RFU) prolonged to the beginning of April 2021, relative to the peak of 0.32 RFU only during January 2020. This suggests poorer water quality during the 2021 rainy period, as well as a gradual increase in the contribution (~doubling) of cyanobacterial biomass from 2019 (≤ 0.2 RFU) to 2021 (≤ 0.4 RFU). Meanwhile, the ratio of cyanobacteria biomass to total phytoplankton biomass dropped slightly from 0.15/0.5 = 30% in 2019 to 0.2/1.0 = 20% in 2021. However, we need the data for 3 full years to confirm this calculation. A ratio of 20-30% is quite high. It suggests keeping a high and permanent vigilance, because if conditions become favorable, cyanobacteria can quickly proliferate and generate blooms that can become recurrent every year.  The damage (mortality of fish, frogs and waterfowl) would be irreparable and repeated in the long term, since cyanobacterial blooms would be impossible to control in Lake Titicaca. In fact, Lake Titicaca has no herbivorous organisms (neither fish nor zooplankton) capable of grazing them effectively, nor wastewater treatment plants (WWTPs) capable of retaining the nutrients used by the algae for their growth. Even if Totora are natural biological filters of nutrients and pollutants, alone in the absence of WWTPs, they are not capable of controlling the biological discharges of a population of 1.2 million inhabitants of the urban area of El Alto. To have beneficial effects, Totora (and wetlands in general) must be associated as a complementary treatment to WWTPs, as is universally practiced throughout the world.

Figure 11 - Temporal evolution of fluorescent dissolved organic matter (fDOM) concentration at 1 m depth. During the dry period it varied between 0.9-1.5 RFU. It increased slightly during the rainy period from 1.7 RFU in February 2020 to 2.1 RFU in November 2020. This increase in February and November resulted from organic matter inputs from the rivers, particularly from the Katari watershed, via the Cohana and Sehuenca rivers joining in the Cumana Lagoon, which flows into Cumana Bay. Counter-intuitively to what was thought about the pandemic and quarantine period in 2020, the concentration of organic matter (fDOM) was much higher in 2020 compared to the 2nd half of 2019 and the 1st half of 2021. In fact, the population did not stop contributing to the organic pollution of rivers and lakes. Contrary to what the press wanted to convince us, no significant improvement in the water quality of the Lago Menor was observed in 2020. Even if the discharges of organic matter were completely stopped in 2020 (impossible situation), the organic matter already deposited for decades with the polluted sediments would not have disappeared in one year with the action (= mineralization) of the microbial loop. In conclusion, it is essential to maintain permanent long-term monitoring, since without the implementation of new WWTPs, there are no mechanisms to eliminate such massive discharges of nutrients and organic matter that benefit microalgae blooms. 

Figure 1 - The data presented here correspond exclusively to surface data at 1.0 m depth (range 0.8-1.2 m). The acquisition frequency is 30 min. Every 2 hours, the profiler (guincho) vertically profiles the EXO2 probe from 1 m to 10 m. It performs a measurement every meter, at the end of the 1 minute dwell time (to stabilize the conditions for the sensors). Each down and up cycle takes ~15 min, so it does not affect the next 1 m measurement. However, we have observed numerous occasions when the EXO2 sonde stays several days (most often 1-2 days, exceptionally up to 5 days) at 10 m**.

**NOTE: This intriguing situation was clarified on 28/04/22 when the divers of CIBA (Centro de Instrucción de Buceo en Altura, Tiquina) replaced all the mooring lines of the buoy, worn by 3 years of storms and waves, about to break, including some shackles badly worn by friction. The GoPro camera video revealed long stretches of damaged and abandoned fishing nets wrapped around the concrete blocks and bottom mooring anchors. Therefore, one can imagine that the EXO2 sonde gets caught in these nets and takes a long time to free itself and ascend.

Although we can detect these situations remotely on the LoggerNet (Campbell Scientific Inc.) screen of our DELL laptop, we do not always have the time, the service vehicle or the budget to drive to the lake, rent the boat, hoping to solve the problem by manually lifting the sonde. Then, we just have to let the swell and the Buoy's movements free the probe from these nets. In fact, the winch of the profiler has a safety device to avoid forcing the electric motor in case of excessive resistance of the cable.

The presence of these abandoned nets reveals the intense artisanal fishing activity in the Buoy area. Of course, it would be ideal if fishermen would retrieve their damaged nets instead of letting them drift and thus pose a danger to the outboard motor propellers, the buoy sounder, and especially to the fish communities (considering that frogs are scarce), as they continue to fish and kill fish uselessly. Obviously, it is not always easy to avoid losing nets. However, environmental education (e.g. with a primer) in this regard could reduce this problem.

Figure 2a - Water temperature evolution at 1 m depth (date format mm/dd/yy). This graph illustrates the effect of the seasonal cycle of solar radiation on water surface temperature. The sinusoidal curve has a large amplitude with maxima (≤ 18 ºC) in summers (November to February) and minima (≥ 10 ºC) in winters (June to August). At first glance the oscillation is regular with no apparent difference between years. Blanks in the data correspond to periods when the EXO2 sonde remained at 10 m depth (corresponding data were removed). Mean +/- standard error = 14.247 +/- 0.009 ºC.

Figure 2b - Comparison of seasonal temperature evolution between years 2019 (blue), 2020 (red), 2021 (green) and 2022 (purple) (m/dd date format). The seasonal patterns are similar between years and overlap. Perhaps, the water warmed slightly faster and more (~+0.5 °C) in September-October 2021. However, it does not seem significant.

Figure 3a - Evolution of specific conductivity. In 2019-2020 it started around 1,510 µS/cm. It increased in the rainy seasons of 2020 (≥ 1,550 µS/cm) and 2021 (≥ 1,650 µS/cm) and decreased in the dry season of 2021 (≤ 1,490 µS/cm), which may seem counter-intuitive. Throughout the study period it has a tendency to increase, although not statistically significant. Mean +/- standard error = 1526.584 +/- 0.357 µS/cm.

Figure 3b - Comparison of the seasonal evolution of specific Sp conductivity between 2019, 2020, 2021 and 2022. What jumps out is: a) Lower conductivity in January-May 2020; b) Higher conductivity in February (≤ 1,570 µS/cm) and December 2021 (≤ 1,650 µS/cm); and c) The highest conductivity in January 2022 (≤ 1,630 µS/cm).

Figure 4 - Comparison of seasonal pH evolution between 2019, 2020, 2021 and 2022. What jumps out is: a) EYE: 2019 (blue) and 2021 (green) values are not reliable; b) pHs are high (> 8.6), with a tendency to increase from January (8.7-9.0) to December (9.2-9.3) in 2020, in case it is not a sensor drift.

Figure 5a - Evolution of the oxidation-reduction potential, ORP. As the ORP sensor is integrated to the pH sensor, the measurements were reliable until September 2020, around +240 mV. Average +/- standard error = 241.977 +/- 0.445 mV. 

Figure 5b - Comparison of the seasonal evolution of ORP between 2019, 2020, 2021 and 2022. What jumps out is: a) EYE: the values from January to May 2021 (green; from June onwards unreliable) and January 2022 (purple; from February onwards unreliable) are the highest (280-400 mV). b) The 2019 (blue) and 2020 (red) values are relatively stable between 280 and 200 mV, at the lower limit of healthy waters. c) Two periods of higher values are observed: December 2021 (300-550 mV) and January 2022 (300-400 mV). As it is a sensor that integrates pH (defective) and ORP measurements, there is a doubt that the ORP measurement may not be very reliable (as in 2021).

Figure 6a - Percent saturation in dissolved oxygen fluctuated between ≤ 70% (December 2021) and ≥ 115% (January 2020). As a trend it is higher during the dry period (June-July) and lower at the end of the dry period (November-December). It is highlighted that it remained high (≤ 110%) throughout the confinement period (March-September 2020), due to the good water quality in the absence of most of the inputs from polluting human activities, low turbidity, strong insolation (low cloudiness), intense vertical mixing during the windy period (July-June), among others. Average +/- standard error = 89.248 +/- 0.054 %.

Table 1 - Calibration values for various atmospheric pressures and altitudes. Source: XYLEM Technical instructions: https://www.ysi.com/File%20Library/Documents/Technical%20Notes/DO-Oxygen-Solubility-Table.pdf 

Figure 6b - Comparison of the seasonal evolution of dissolved oxygen saturation (DO, %) between 2019, 2020, 2021 and 2022. What jumps out is: a) Better DO% saturation in January 2020 (> 110%), relative to the other years; b) Higher saturation (up to > 100%) than the other years, during confinement (March-September 2020); and c) Bad saturations in December 2021 (≤ 70%).

Figure 7a - The dissolved oxygen concentration follows the behavior of the percent saturation in dissolved oxygen in Fig. 6. Mean +/- standard error = 9.126 +/- 0.006 mg/L.

Figure 7b - Comparison of the seasonal evolution of dissolved oxygen concentration (DO, mg/L) between the years 2019, 2020, 2021 and 2022. What jumps out is: a) Logically, the DO patterns (mg/L) are similar to DO patterns (%) (see Fig. 6b).

Figure 8a - Evolution of turbidity. Average +/- standard error = 0.226 +/- 0.002 NTU. 

Figure 8b - Comparison of the seasonal evolution of turbidity between 2019, 2020, 2021 and 2022. What jumps out is: a) The lowest values during confinement (March-June 2020); b) The exponential increase in turbidity from October to November 2021; and c) The maintenance of higher values (up to 0.6 NTU) in April 2022.

Figure 9a - Evolution of chlorophyll-a concentration. Mean +/- standard error = 0.875 +/- 0.003 RFU.

Figure 9b - Comparison of the seasonal evolution of chlorophyll-a concentration between 2019, 2020, 2021 and 2022. What jumps out is: a) In relation to the other years, the confinement period (March-October 2020, red) exhibits the lowest chlorophyll-a concentrations; b) The 2021 rainy season (September-December, green) exhibits the highest chlorophyll-a concentrations.

Figure 10a - Evolution of phycocyanin concentration. Average +/- standard error = 0.208 +/- 0.001 RFU.

Figure 10b - Comparison of the seasonal evolution of phycocyanin concentration between 2019, 2020, 2021 and 2022. What stands out is: a) The increase in concentration during the end of 2021 and the beginning of 2022, i.e. the transition period at the end of the dry season (September-October) and the rainy season (November-January), with a doubling of values.

Figure 11a - Evolution of fluorescent dissolved organic matter (= chromophoric). Mean +/- standard error = 1.868 +/- 0.007 RFU.

Figure 11b - Comparison of the seasonal evolution of fDOM concentration between 2019, 2020, 2021 and 2022. What jumps out is in relation to the average value (1.3 RFU): a) The increase of fDOM with maximum values in March-May 2021 and January 2022 (up to 4.8 RFU) in relation to the other years, and b) Its low values from July to December 2019, from January to December 2020, in June-July 2021, and in April 2022. It could result from the resumption of gravel extraction activities (in January 2021) in the Seke and Seco river quarries, heavy rains and/or the diversion of the Katari river towards Chojasivi, with heavy flooding of crops in Cohana Bay, all of which contribute to diffuse inputs of nutrients and dissolved organic matter (DOM). This clearly initiates a jump in the water quality status of the northern region of Lesser Lake. This should be carefully observed, because it may favor the proliferation of certain groups of micro-algae, in particular the harmful cyanobacteria.