We present long‐term changes in Norwegian lake water quality across regional gradients in atmospheric pollution, air temperature, hydrology, and vegetation using (a) a national representative lake survey carried out in 1995 and 2019 (ThousandLakes), and (b) an annual lake survey from acid‐sensitive catchments (78 lakes, TrendLakes) from 1990 to 2020. Our analysis encompasses all major chemical constituents, for example, anions and cations, dissolved organic matter (DOM), nutrients, iron (Fe), and silicate (SiO₂). During these decades, environmental changes included declines in sulfur (S) and nitrogen (N) deposition, climate warming, and increase in forest biomass. Strong chemical recovery from acidification is found, attributed to large reductions in atmospheric deposition, moderated by catchment processing from land use and climate change. Browning counteracted chemical recovery in some regions, while Ca increased unexpectedly. We suggest that increased weathering, from enhanced terrestrial productivity, is an important driver of increased Ca—substantiated by widespread, substantial increases in SiO₂. Light‐ and nutrient‐limitation has become more prevalent, indicated by higher DOM, lower nitrate (NO₃), and lower NO₃to total phosphorous ratios. Declines in lake NO₃ occurred independently of N deposition, suggesting increased catchment N retention, possibly from increased terrestrial productivity. We conclude that decreased air pollution continues to be a dominant driver of long‐term trends in lake chemistry, but climate‐induced increase in terrestrial weathering processes, governed by increased biomass, is likely to have an increasing impact on future lake acidity, nutrient, and light status, that may cascade along the aquatic continuum from rivers to the coast.

Freshwater management is challenging, and advance warning that poor water quality was likely, a season ahead, could allow for preventative measures to be put in place. To this end, we developed a Bayesian network (BN) for seasonal lake water quality prediction. BNs have become popular in recent years, but the vast majority are discrete. Here, we developed a Gaussian Bayesian network (GBN), a simple class of continuous BN. The aim was to forecast, in spring, mean total phosphorus (TP) and chlorophyll a (chl a) concentration, mean water colour, and maximum cyanobacteria biovolume for the upcoming growing season (May–October) in Vansjø, a shallow nutrient-rich lake in southeastern Norway. To develop the model, we first identified controls on interannual variability in seasonally aggregated water quality. These variables were then included in a GBN, and conditional probability densities were fit using observations (≤39 years). GBN predictions had R² values of 0.37 (chl a) to 0.75 (colour) and classification errors of 32 % (TP) to 17 % (cyanobacteria). For all but lake colour, including weather variables did not improve the predictive performance (assessed through cross-validation). Overall, we found the GBN approach to be well suited to seasonal water quality forecasting. It was straightforward to produce probabilistic predictions, including the probability of exceeding management-relevant thresholds. The GBN could be sensibly parameterised using only the observed data, despite the small dataset. Developing a comparable discrete BN was much more subjective and time-consuming. Although low interannual variability and high temporal autocorrelation in the study lake meant the GBN performed only slightly better than a seasonal naïve forecast (where the forecasted value is simply the value observed the previous growing season), we believe that the forecasting approach presented here could be particularly useful in areas with higher sensitivity to catchment nutrient delivery and seasonal climate and for forecasting at shorter (daily or monthly) timescales. Despite the parametric constraints of GBNs, their simplicity, together with the relative accessibility of BN software with GBN handling, means they are a good first choice for BN development with continuous variables.

Advance warning of seasonal conditions has the potential to assist water management in planning and risk mitigation, with large potential social, economic, and ecological benefits. In this study, we explore the value of seasonal forecasting for decision-making at five case study sites located in extratropical regions. The forecasting tools used integrate seasonal climate model forecasts with freshwater impact models of catchment hydrology, lake conditions (temperature, water level, chemistry, and ecology), and fish migration timing and were co-developed together with water managers. To explore the decision-making value of forecasts, we carried out a qualitative assessment of (1) how useful forecasts would have been for a problematic past season and (2) the relevance of any windows of opportunity (seasons and variables where forecasts are thought to perform well) for management. Overall, water managers were optimistic about the potential for improved decision-making and identified actions that could be taken based on forecasts. However, there was often a mismatch between those variables that could best be predicted and those which would be most useful for management. Reductions in forecast uncertainty and a need to develop practical, hands-on experience were identified as key requirements before forecasts would be used in operational decision-making. Seasonal climate forecasts provided little added value to freshwater forecasts in these extratropical study sites, and we discuss the conditions under which seasonal climate forecasts with only limited skill are most likely to be worth incorporating into freshwater forecasting workflows.