Are more data always better? - Machine learning forecasting of algae based on long-term observations

Aug 3, 2026

Upload by Ezra

Bloom-forming algae significantly hinder water management by impairing ecosystem services and creating health risks for humans and animals. Reliable short-term algae forecasts offer early warnings and support timely mitigation. This study evaluates how much training data is needed for machine learning models to forecast chlorophyll-a two weeks ahead using 30 years of fortnightly measurements from a UK lake. Results show performance gains up to about four to five years, with diminishing returns and feature-selection benefits.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages13

Reading Time33 min

AccessFree to download

Abstract

Introduction

Impacts of algal blooms

Management and mitigation approaches

Need for early-warning forecasting

Machine learning approaches for algae forecasting

Journal of Environmental Management 373 (2025) 123478 Contents lists available at ScienceDirect Journal of Environmental Management journal [homepage: www.elsevier.com/locate/jenvman](homepage: www.elsevier.com/locate/jenvman) | Research article
Are more data always better?– Machine learning forecasting of algae based on long-term observations
D. Atton Beckmanna,*, M. Werther b, E.B. Mackay c, E. Spyrakosa, P. Hunter a,d, I.D. Jones a
a Biological and Environmental Sciences, School of Natural Sciences, University of Stirling, Stirling, United Kingdom
b Swiss Federal Institute of Aquatic Science and Technology, Department of Surface Waters-Research and Management, Dübendorf, Switzerland c UK Centre for Ecology and Hydrology, Lancaster Environment Centre, Lancaster, LA1 4AP, United Kingdom
d Scotland’s International Environment Centre, School of Natural Sciences, University of Stirling, Stirling, United Kingdom | | | | --- | --- | --- | | A R T I C L E I N F O | | A B S T R A C T | | Handling editor: Lixiao Zhang | | Bloom-forming algae present a unique challenge to water managers as they can significantly impair provision of important ecosystem services and cause health risks to humans and animals. Consequently, effective short-term algae forecasts are important as they provide early warnings and enable implementation of mitigation strategies. In this context, machine learning (ML) emerges as a promising forecasting tool. However, the performance of ML models is heavily dependent on the availability of appropriate training data. Consequently, it is essential to determine the volume of data necessary to develop reliable ML forecasts. Understanding this will guide future monitoring strategies, optimize resource allocation, and set realistic expectations for management outcomes. In this study, we used 30 years of fortnightly measurements of 13 different parameters from a lake in the English Lake District (UK) to examine the impact of training data duration on the performance of ML models for forecasting chlorophyll-a two weeks in advance. Once training data availability exceeded four years, a Random Forest model was found to consistently outperform naive benchmarks (mean absolute percentage error 16.4 % lower than the best-performing benchmark). With more than 5 years of training data, model performance generally continued to improve, but with diminishing returns. Furthermore, it was found that equivalent and, in some cases, better performance could be achieved by only using a subset of the most important input features. Additionally, it was found that reducing the sampling frequency had negative impacts on performance, both due to the reduced number of training observations available, and increased forecast horizon. Our findings demonstrate that for lakes ecologically similar to the study site, a consistent and regular sampling programme focused on monitoring a limited number of key parameters can provide sufficient observations for generating short-term algae forecasts after approximately five years of data collection. Importantly, this result provides justification for the initiation of new monitoring programmes for sites where algal blooms are a concern, and suggests that there are likely many pre-existing monitoring datasets which would be suitable for training algae forecast models. | | Keywords:
Algal blooms Cyanobacteria Forecasting Freshwater Early warning Machine learning | | | 1. Introduction Excessive growths of algae in freshwater, commonly referred to as algal blooms, compromise the safety of drinking water sources (Brooks et al., 2016; Igwaran et al., 2024), endanger recreational water activities (Carvalho et al., 2013; Wolf et al., 2017), and threaten the stability and diversity of aquatic ecosystems (Amorim and Moura, 2021; Dolah et al., 2001). Blooms can also impose significant economic impacts, affecting tourism, the fish industry, and even property prices (Hamilton et al., 201

FAQs

Why are short-term algae forecasts important for water management?

Bloom-forming algae can impair ecosystem services and create health risks. Short-term forecasts provide early warnings, enabling timely mitigation actions.

How does the study examine the effect of training data duration on ML forecasting?

It uses 30 years of fortnightly measurements across 13 parameters from a UK lake to train machine learning models and evaluate forecast performance for chlorophyll-a two weeks ahead.

What training data amount was found to improve forecast reliability, and how did returns change over time?

When training data exceeded four years, a Random Forest model consistently outperformed naive benchmarks. Performance generally improved with more than five years of data, but benefits showed diminishing returns.

Aug 3, 2026

Upload by Ezra