The water forecast: predicting safe bathing water
In summer, blue-green algae grow in many Dutch lakes. Water managers can issue a warning or a negative bathing advice based on periodic sampling. However, the amount of blue-green algae varies greatly depending on the time and place, partly due to weather conditions. As a result, despite a warning or negative advice, the water in between sampling times may still be safe, or unsafe despite no warning or negative advice having been issued. Can the Algae Radar help make bathing advice more reliable?
The current practice is that of the Blue-Green Algae Protocol 2020
[1]. This protocol provides guidelines for drawing up bathing advice: the water manager measures blue-green algae concentrations on a weekly or bi-weekly basis and bases bathing advice on these: warning (12–75 µg/L), negative bathing advice (>75 µg/L). The problem with this method is that the data are never up to date. It would be better if blue-green algae concentrations could be predicted. To that end, the Algae Radar is being developed (see box). The Algae Radar links periodic measurements of blue-green algae to historical average weather data from previous years.
Algae Radar
The Algae Radar is still under development. It is a data-driven model from Deltares that produces blue-green algae forecasts based on blue-green algae measurements and meteorological data (wind speed, air temperature and solar radiation[2]. Previously, Algae Radar forecasts were made exclusively based on weather averages from previous years. This project uses current weather forecasts for the first time. In this way, the blue-green algae forecast is given a level of uncertainty linked to the uncertainty in the weather forecast – the blue-green algae plume. The expectations visualise the extent to which risk levels (warning or negative bathing advice) are exceeded (Figure 1).
Sometimes, blue-green algae near the banks can cause additional nuisance by forming floating layers. The latter results from the movement of water and leads to extremely high concentrations. This phenomenon cannot be predicted by the Algae Radar, since water movement is not included in the model. However, it is possible to take measures, such as flushing or peroxide treatment, based on the Algae Radar forecast, even before the nuisance starts. The Algae Radar can also be used to test the effectiveness of measures.
Deltares, the Royal Netherlands Meteorological Institute (KNMI) and the Netherlands Institute of Ecology (NIOO-KNAW) are working together to improve the Algae Radar. In this project, we investigate whether the Algae Radar can better predict blue-green algae concentrations by using current weather forecast data rather than
historical weather data. This article presents the first results of such a blue-green algae forecast made on the basis of the weather forecast, or 'water forecast'.
To check whether a water forecast indeed better reflects the reality in the field, the model results are compared with high-frequency and other measurements in the field (Paterswoldsemeer lake).
In the field
To this end, we conducted high-frequency measurements of blue-green algae and other relevant parameters in two Dutch lakes during the bathing season from 1 May to 1 October 2022: the Paterswoldsemeer lake near Groningen and the Wylerbergmeer lake near Nijmegen. High-frequency measurements in Paterswoldsemeer were compared with sampling by the Noorderzijlvest Water Board and the Algae Radar blue-green algae forecasting model.
For the high-frequency measurements in this project, NIOO-KNAW is using two new measuring buoys. These determine concentrations of blue-green algae and total chlorophyll every 10 minutes at a depth of 50 centimetres based on measurements of fluorescence, oxygen, acidity and conductivity. The buoys are powered by a solar panel and send hourly data to a server, from where the data can be analysed. They thus provide a near real-time picture of the various parameters and thus insight into the daily dynamics of blue-green algae.
Weather plume
The KNMI has access to ensemble weather forecasts of temperatures and other meteorological quantities – the 'weather plume'[3]. These are 51 separate forecasts, which together give a good idea of the possible weather over the next two weeks. The ensemble weather forecast thus gives the uncertainty of that forecast in addition to the forecast itself. The ensemble weather forecast and blue-green algae measurements (standardised according to the Blue-Green Algae Protocol) are entered into the Algae Radar.
Advice based on buoy measurements and models
The buoy measurements show daily dynamics between mid-August and mid-September 2022 (Figure 1). Immediately after the start of measurements, the blue-green algae concentration increased for a period of about one week, leading to a negative bathing advice (Figure 1A, B). After this period, concentrations decreased and, based on the buoy measurements, the advice would change to a warning around the beginning of September. The Algae Radar's blue-green algae forecast model showed a similar pattern based on the five-day weather forecast, but now the concentration remained high until just before the end of the measured period, resulting in a negative bathing advice for almost the entire period. The buoy measurements also show some variation throughout the day, where in some cases the lowest measured values would arrive at a different advice than if the average were used.
Figure 1. Bathing advice according to the standards of the Blue-Green Algae Protocol 2020[5], based on high-frequency buoy measurements (A, B) and on (five-day) Algae Radar blue-green algae forecasts (C, D) for the Paterswoldsemeer lake from 15 August to 13 September 2022. The continuous black lines (B,D) give daily averages, with dotted lines indicating the minimum and maximum. The grey bars show the blue-green algae concentrations as reported by the water manager[5].
Model results
The results from the Algae Radar model remained above the level of a negative bathing advice for almost the entire period from 15 August to 1 September. As mentioned earlier, the Algae Radar calculates blue-green algae concentrations based on meteorology and blue-green algae measurements. For the latter, the water board's monthly measurements were used. The most recent measurements weigh most heavily in determining future concentrations. Figure 1D shows that the mid-August reading is high (grey bar). Combined with weather conditions favourable to blue-green algae (high temperature, high irradiance and low wind speeds), this leads to model results remaining high throughout the period until mid-September. After all, the model adjusts expectations only once new blue-green algae measurement data come in, and they did not come in during this period. In the future, it will be possible to use data from the measurement buoys. However, this will not be possible for another few years, because calibrating the Algae Radar has so far required a historical measurement series of at least six years. This period can be shortened by measuring more frequently with the measurement buoy.
Added value of buoy measurements
The buoy measurements show that the concentration of blue-green algae can vary more than is visible with less frequent field sampling. In this case, this leads to an overestimation of the risk in the second half of the measurement period. The Blue-Green Algae Protocol 2020 prescribes weekly or bi-weekly measurements. If followed, it is plausible that, for the second period as well, the advice issued would better reflect the buoy measurements and thus the actual situation. Therefore, for the limited dynamics during such a short measurement period, standard monitoring will be sufficient. It is possible that daily dynamics will increase during the initial phase of blue-green algae blooms and/or stronger variation in weather. In that case, buoy measurements can help by providing a remote, direct view of the situation at a bathing site. In addition, targeted monitoring based on such buoy measurements could even be possible, with the possibility to take additional field measurements any time the buoy changes its advice.
Water forecast and climate change
Climate change is causing temperatures to rise. This is also expected to increase nuisance from blue-green algae. As a result, bathing sites may be closed more often and for longer. At the same time, there is a greater need to cool down. Water managers can use water forecasts to more accurately, and in advance, determine the risk of standards being exceeded. This will allow more effective and risk-driven sampling, i.e. with less cost and effort, allow any measures to be taken earlier and potentially allow bathing sites to remain open more often.
Follow-up work
If the water forecast is to be used in practice, the buoy measurements could be integrated into the forecast model to increase short-term certainty. In addition, the reliability of the forecast can be tested for different conditions, such as the depth and size of a body of water or the amount of blue-green algae. Finally, a dashboard that directly displays measured data and model forecasts would be useful. This will allow us to view blue-green algae concentrations almost in real time, as well as the forecast for the coming days. As a result, targeted monitoring will be possible, for example by not sampling if the situation is expected to remain unchanged. For an even better expectation of risks, the expectation model should be extended to include toxin concentrations. After all, the health risk is due to the toxins excreted by the blue-green algae[4].
Thanks
This project was funded from the Digital Government 2021 Innovation Budget under the name Voorspellen Veilig Zwemwater (Predicting Safe Bathing Water), project number 1500025049091. The project was carried out under the supervision of stakeholders from Leisurelands, the Rivierenland Water Board, the Noorderzijlvest Water Board and the Dutch Association of Regional Water Authorities.
Miguel Dionisio Pires
(Deltares; Aeres University of Applied Sciences Almere)
Peter Siegmund
(Royal Netherlands Meteorological Institute)
Sven Teurlincx
(Netherlands Institute of Ecology)
Dedmer van de Waal
(Netherlands Institute of Ecology; University of Amsterdam)
Marc Weeber
(Deltares)
Background picture:
The Water Forecast: information board at the Wylerbergmeer lake
Summary
During blue-green algae blooms in summer, bathing water managers may issue a warning or negative bathing advice. However, the amount of blue-green algae varies greatly depending on the time and place, partly due to weather conditions. As a result, the bathing advice may temporarily not match the actual water quality in between sampling sessions. Based on an expectation of blue-green algae, sampling can be risk-driven and bathing waters can reopen earlier after nuisance. This project has taken a first step for such an expectation. During the 2022 bathing season, high-frequency measurements of blue-green algae and other relevant parameters were taken in two lakes in the Netherlands. The measurements were compared with a forecast using the Algae Radar model – a 'water forecast' – based on low-frequency measurements by the water board and detailed weather forecasts.
Sources
1. Sketch, F.M., et al. (2020), Blauwalgenprotocol 2020 (Blue-Green Algae Protocol 2020), RIVM letter report 2020-0107.
2. Deltares: Algenradar voorspelt algenbloei (Algae Radar predicts algal blooms). https://cms.deltares.nl/assets/common/downloads/Flyer-Algenradar-voorspelt-algenbloei-1.pdf
3. KNMI: Weer- en klimaatpluim en Expertpluim (Weather and climate plume and Expert plume), https://www.knmi.nl/nederland-nu/weer/waarschuwingen-en-verwachtingen/weer-en-klimaatpluim
4. Van de Waal, D.B., et al. (2023), Risicobeoordeling van blauwalgen: snel, nauwkeurig of beide? (Blue-green algae risk assessment: Fast, accurate or both?). Water Matters no. 17, December 2023.
^ Back to start