The water quality of surface streams is impacted by various environmental (soil texture, precipitation, and local topography) and anthropogenic (fertilizer and manure deposition) factors of the upstream catchment. Knowledge of relationships between the water quality and the catchment-wide characteristics is of high importance for outlining critical areas for interventions, e.g., nature-based solutions for nutrient capture.
Various modelling techniques have been implemented to gain insights into the catchment characteristics and the corresponding nutrient concentrations. The use of machine learning methods for this purpose has increased due to the relaxed requirements of the input data as well as increasingly ubiquitous spatial environmental datasets. However, machine learning models are not spatially-aware by default. Recently, various methods have been proposed to account for spatial dependency across multiple modelling stages.
We employ the Random Forest supervised machine learning algorithm to model nutrient (nitrogen and phosphorous) concentrations on a point scale. We use national level monitoring data of the Baltic countries between the years 2017-2023 with varying number of observations per site, averaged over the study years. Environmental characteristics (topography, land use, climate, soil properties) describing the corresponding upstream catchment area are used as the explanatory features. As the catchments extend beyond the borders of the Baltics, we use various global datasets for feature creation (e.g., ERA5-Land, SoilGrids). In addition, we apply spatial machine learning methods and assess their applicability for catchment-based modelling. Lastly, we employ explainable AI methods, namely SHapley Additive exPlanations and Partial dependency plots, to validate if our model’s revealed relationships correspond to the domain knowledge.
How to cite: Jemeljanova, M., Virro, H., Annusver, M., Kmoch, A., and Uuemaa, E.: Nutrient concentration modelling in the Baltic countries using spatial machine learning, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-8238, https://doi.org/10.5194/egusphere-egu25-8238, 2025.
Phosphate (PO₄) is often the limiting nutrient driving eutrophication and algal blooms in freshwater ecosystems. Accurate estimation of PO₄ concentrations is crucial for assessing the impacts of agriculture and urban emissions on water bodies and evaluating the effectiveness of catchment mitigation measures. However, measuring PO₄, especially at low concentrations, is technically complex, and modelling its dynamics is challenging due to the interplay between emissions, absorption, and transport processes in the natural environment. This study develops a Random Forest model to predict PO₄ concentrations in Irish water bodies, using catchment descriptors related to climate, land use, geology, topography, and anthropogenic activities. The model achieves an R² of 0.35 and an RMSE of 0.03 mg/L on an independent validation dataset, demonstrating moderate predictive accuracy. The developed model is applied to current and future climate and land use change scenarios to evaluate the influence of catchment-level mitigation measures. The findings highlights the significant roles of soil texture and permeability in influencing downstream PO₄ concentrations. Moreover, descriptors which are associated with low-intensity land use, such as peatlands and forests, were identified as having a positive effect in reducing PO₄ concentrations in water bodies. This underscoring the importance of sustainable land management practices in maintaining healthier ecosystems.
How to cite: O'Loughlin, F., Khan, S., and Mockler, E.: The Role of Catchment Characteristics in Phosphate Emissions to Downstream Waterbodies, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-9173, https://doi.org/10.5194/egusphere-egu25-9173, 2025.
Predicting the concentration and identifying the source of phosphorus in aquatic systems are essential for ecosystem health. This study tackled two primary challenges: the intricate biogeochemical cycle of phosphorus, which hinders the accuracy of process-based models, and the time-intensive, resource-demanding nature of experimental and model-based phosphorus tracing methods. We adopted a novel attention physics-guided spatiotemporal graph convolutional neural network, which employs convective diffusion equations to constrain deep learning training for more accurate spatiotemporal multi-node total phosphorus (TP) predictions, and is coupled with an attention-based interpretability method to trace pollution sources. In application to the Taihu Lake Basin (China), this model enhanced TP concentration prediction accuracy by 7.1%–12.3% compared with baseline models. It also effectively identified and quantified the primary pollution source in Gehu Lake under varying seasonal and hydraulic engineering conditions. Examination of the microscale TP migration process revealed an equilibrium mode between TP concentration dilution and sediment disturbance–release under specific river velocity, with an equilibrium velocity of 0.19 m/s. This study underscores the critical role of hydrodynamics, shaped by hydraulic engineering and hydrological variability, in influencing pollutant migration and transformation within tidal river networks, thereby offering new insights into phosphorus prediction and source tracing in complex habitats.
How to cite: Yao, J. and Ruan, X.: Explainable deep learning for dual goals: Predicting total phosphorus concentrations and identifying pollution sources in the Taihu Lake Basin, a tidal river network, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-4892, https://doi.org/10.5194/egusphere-egu25-4892, 2025.
A developed machine learning framework for predicting ammonium (NH₄⁺) levels in River Lee, London, is presented in this paper. We use state-of-the-art algorithms such as the Temporal Fusion Transformer (TFT), Random Forest (RF), and Extreme Gradient Boosting (XGBoost) on a large dataset that includes temperature, turbidity, chlorophyll, dissolved oxygen, conductivity, and pH. By clarifying the intricate relationships between environmental variables and ammonium levels, these models greatly improve forecast accuracy. Using the TFT model for multi-horizon forecasting is one of our research's unique features. High accuracy and interpretability in hydrological predictions are made possible by this model's skilful integration of convolutional elements with an attention mechanism. It solves a crucial problem in environmental modelling by skilfully managing short-term variations while being resilient over longer periods of time. Adaptability and resilience are combined in our dual-scale method, which works well for both short- and long-term environmental projections. In particular, XGBoost performs exceptionally well in monthly forecasts up to 12 months with a noticeably low RMSE, while the RF model exhibits exceptional long-term forecasting capabilities, attaining an R2 of 0.97 and an RMSE of 0.18 over 1095 days. TFT performs best in short-term projections, but data granularity limits its ability to perform well in longer-term situations. These revelations highlight how urgently proactive water management techniques are needed to reduce hazards like hypoxia and possible ecological effects. In the end, our research offers resource managers vital assistance in tackling issues pertaining to ammonium toxicity and ecological health.
How to cite: Ali, A. and Ahmed, A.: AI prediction of ammonium levels in rivers using machine learning, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-10930, https://doi.org/10.5194/egusphere-egu25-10930, 2025.
Predicting dissolved oxygen (DO) levels in river ecosystems—particularly in River Lee, London—is crucial to maintaining aquatic life and water quality. Using daily data, this work presents machine learning models that can forecast DO levels over a variety of periods, from short (7 days) to long (365 days). We enhanced the capacity of long-term DO forecasting by utilizing models such as Temporal Fusion Transformer (TFT), Informer, Long Short-Term Memory (LSTM), and Gated Recurrent Units (GRU). With an RMSE of 0.04 and an R2 of 0.09 at the 365-day horizon, the Informer model performs well in managing long-term dependencies. On the other hand, although the TFT model consistently performs well throughout a range of time periods, the LSTM and GRU models' accuracy decreases for forecasts longer than 90 days. Furthermore, DO levels are greatly influenced by environmental factors such as pH, chlorophyll, turbidity, temperature, conductivity, and river velocity. Environmental organisations can develop proactive water management plans and prevent problems like river hypoxia thanks to the enhanced performance of models like the Informer and TFT. These results highlight how cutting-edge machine learning methods can help ensure the long-term viability of river ecosystems.
Keywords: River streamflow; LSTM; GRU; TFT; Informer; Water Quality
How to cite: Ahmed, A. and Ali, A.: Long-Term Forecasting of Dissolved Oxygen in Rivers Using Machine Learning Models, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-9554, https://doi.org/10.5194/egusphere-egu25-9554, 2025.
Rivers play a crucial role in in global matter cycling and energy flow, contributing significantly to biogeochemical
cycles and the development of human civilization. Reservoirs, as prevalent artificial water bodies, modify river
flow and impact energy and environmental dynamics. These reservoirs can directly affect riverine ecosystems by
retaining algal materials, thereby altering Chl-a concentrations in downstream water bodies. Nevertheless, the
mechanisms by which reservoirs influence Chl-a concentrations in rivers remain poorly understood. This study
utilized Landsat 8/9 images and in-situ measurements from the Pearl River to develop a machine learning model
and generate a Chl-a concentration dataset spanning 2013-2022. We also examined the mechanisms through which
reservoirs and the natural environmental factors affect Chl-a concentrations by regulating the Pearl River. The
findings indicate that anthropogenic factors, primarily the construction of reservoirs and dams, play a significant
role in shaping the spatial distribution of riverine Chl-a concentrations along the Pearl River. As the river traverses
reservoirs in the upper and middle reaches, Chl-a concentrations in both the mainstem and tributary sections
exhibit a distinct decrease. The highest Chl-a concentrations were observed in the headwaters of the Xijiang River,
followed by a decline in the midstream, and a subsequent increase downstream. It also revealed that, river Chl-a
levels are consistently lower before entering a reservoir, higher within it, and further decreased after exiting.
Reservoirs, by intercepting and storing upstream sediment and nutrients, allow only a small amounts to pass
through dams into downstream sections, thereby influencing riverine Chl-a concentrations. Furthermore, Chl-a
concentrations in the Pearl River peak during summer and reach their lowest levels in winter, with water
temperature being the dominant driver of seasonal and interannual Chl-a variations (r = 0.88, p < 0.01). Other
environmental factors such as pH, dissolved oxygen (DO), total phosphorus (TP), total nitrogen (TN), and Chl-a
concentrations were found to be positively correlated. Our findings indicate that cascade reservoirs have a more
significant impact on river environmental status. To effectively address river water quality degradation and
maximize the benefits of reservoirs, coordinated water diversion and protective measures between the reservoirs
are required.
How to cite: Li, Z.: Assessing the Impacts of Cascade Reservoirs on Pearl River Environmental Status Using Machine Learning and Satellite-derived Chlorophyll-a Concentrations, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-505, https://doi.org/10.5194/egusphere-egu25-505, 2025.
Water quality modelling at the operational management catchment scale suffers from large epistemic uncertainties with reduced monitoring and uncertain processes that are rapidly changing with the climate. Assessing the effectiveness of many distributed nature-based solutions (NbS) that change hydrological and geochemical processes in the landscape to help mitigate diffuse nutrient pollution can be even more beset by uncertainties and lead instead to a focus on asset-only improvements, and a corresponding loss of multiple benefits associated with NbS.
Rather than taking a complex integrated catchment modelling approach, we focus here on using regulatory, data-based model, SIMCAT, to first understand the least change in diffuse load NbS must deliver to improve WFD status. This shift can be monetised and combined with other co-benefits of NbS including water resource, habitat and carbon estimated here from additional models. This helps identify waterbodies where the least effort on behalf of NbS is required to improve the status, and these areas can be refined by combining with waterbodies with the greatest potential for NbS. This potential has been mapped in the UK delineating areas for potential wetland restoration, woodland planting, ponds and floodplain restoration. These have been combined and the intersected area of these for each of the WFD waterbodies can then help prioritise further, and assessed against water company plans for future asset-improvements. The process results in a multi criteria analysis where we explore trade-offs between different benefits and mixed solutions that include pipeline future asset improvements.
Having collaboratively agreed the weightings in the MCA and agreed the target WFD waterbody catchments where NbS will be most effective, we introduce an additional step of using a novel 10m gridded risk map from the Fieldmouse model, which identifies pixels in the landscape with greatest load and connectivity to the classification point. This is used as a heat map to refine which part of the mapped NbS elements would make the greatest difference and where for instance grant allocation can be focussed. The modelling tool can also quantify the reduction in load, although this can still be quite uncertain, the measures are located where they are likely to make the most difference.
How to cite: Broomby, J., Hankin, B., Wang, C., Champion, H., Maslen, S., and Gerrard, C.: Quantifying and targeting the multiple benefits of nature based solutions at the catchment scale, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-4314, https://doi.org/10.5194/egusphere-egu25-4314, 2025.
Despite long-term regulatory controls on fertiliser management that effectively close and open spreading periods, there are still ongoing stream water quality issues in agricultural catchments. Adjustments to these regulations largely relate to application rate and set-back distances from watercourses at the start of the open period to avoid sudden water quality impacts. Within this regulatory framework and using long-term datasets the aim of this study was to investigate the relative importance of weather, land use and policy effects on stream water quality during the first weeks of the open spreading period. Fortnightly stream water samples were collected over 2009-2023 in twenty-four agricultural sub-catchments of major Northern Ireland rivers. Random Forest Regression models were developed to predict baseline stream water total phosphorus (TP), soluble reactive phosphorus (SRP) and total oxidised nitrogen (TON) concentrations. Results showed that weather and land use were the primary drivers of changes in phosphorus concentrations while land use was the primary driver of changes in TON concentrations. Furthermore, weather was a more important driver of changes in nutrient concentrations in the more intensively farmed sub-catchments. In the less intensive sub-catchments, land use was at least 30% (for TP) to 85% (for TON) more important than the weather and policy predictors for explaining these changes. The study highlights the need to reduce the nutrient source pressure as a more effective step to improve water quality compared to small adjustments to fertiliser spreading protocols. It further supports the need to ensure slurry is spread when weather conditions are appropriate and for policy reviews to account for changes in weather pressures.
How to cite: Fresne, M., Jordan, P., and Cassidy, R.: Relative importance of weather, land use and slurry spreading regulations for stream water quality in agricultural catchments, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-3531, https://doi.org/10.5194/egusphere-egu25-3531, 2025.
Diffuse pollution is a global issue where management, particularly of phosphorus (P) loss from agricultural land to water, must address both source and pathway pressures concurrently as part of effective mitigation. Where this is a widespread issue, policy makers, agri-environmental managers and farmers need a process of prioritisation that places the delivery point for diffuse P to a waterbody into a wider context of risk and maximises the impact of any mitigation for the limited resource available. However, data requirements and lack of a unified method have made this difficult to implement.
This study considers this challenge using field-by-field soil test P monitoring and high-resolution LiDAR runoff risk modelling being developed for all agricultural land in Northern Ireland through the Soil Nutrient Health Scheme. We combine long-term available water quality data for macro- and meso-scale catchments with this unique spatially explicit data set on soil test P and runoff risk (Hydrologically Sensitive Area (HSA)) combinations to rank risk to water quality down to a base unit of a micro-catchment scale (0.02 – 1.6 km2) delineated upslope from each delivery point to a waterbody. This is expressed as a dimensionless Source:Pressure Priority Index (SPPI) which conveys the combined source and pathway risk at a location but without any dimensioned values that would link to soil test P in specific fields and affect confidentiality of field-scale nutrient status information. With an average of 250 delivery points km-2 this approach can filter the highest category SPPI areas to ~1% of those micro-catchments where measures should be targeted first.
This combination and analysis of “big data” provides a whole-landscape risk ranking method for diffuse pollution management that can be directed centrally and rolled out more locally as part of catchment level agri-environmental schemes (AES) and in targeting advisory and extension services. This will ensure a faster route to diffuse pollution reduction and offer resilience as pathway mitigations become vulnerable to weather patterns and runoff responses in a changing climate.
How to cite: Cassidy, R., Service, T., Atcheson, K., Harrison, T., Higgins, A., Farrow, L., Jack, P., and Jordan, P.: Diffuse pollution management in agricultural landscapes – a combined Source:Pathway Priority Index to target advice and resources for impact. , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-19242, https://doi.org/10.5194/egusphere-egu25-19242, 2025.
The South China region is characterized by a monsoon climate with concurrent rainfall and high temperatures, featuring abundant precipitation, dense water networks, and numerous lakes. With the development of agriculture and industry, water pollution has become increasingly severe. Dissolved oxygen (DO), as a crucial indicator for water quality assessment, effectively reflects changes in water quality. In recent years, water quality in aquaculture ponds, rivers, and lakes has gradually improved due to advances in sewage treatment technology and strengthened water quality management. Based on Landsat 8/9 OLI satellite imagery, this study applied Rayleigh reflection atmospheric correction, combined with EMD decomposition and water body indices, to construct a random forest retrieval model for dissolved oxygen (R²=0.90). The study analyzed the spatiotemporal variations of DO concentrations in three typical water bodies across South China from 2013 to 2024. Results showed that: (1) DO concentrations in all three water body types exhibited an increasing trend from 2013 to 2024, with increases of 0.2%, 0.8%, and 2.4% in rivers, aquaculture ponds, and lakes, respectively; (2) In terms of average DO concentration, lakes maintained the highest levels (7.93 mg/L), followed by rivers (7.75 mg/L), while aquaculture ponds showed the lowest levels (7.41 mg/L); (3) Spatially, DO concentrations in rivers decreased gradually from upstream to estuaries, lake centers showed higher concentrations than shoreline areas, and aquaculture ponds demonstrated higher levels in mountainous and upstream river regions compared to lower-latitude coastal areas; (4) Seasonal patterns revealed that DO concentrations in rivers and lakes reached their minimum in summer and maximum in winter, while aquaculture ponds showed an opposite trend.
How to cite: Mao, K. and Yang, X.: Spatiotemporal Characteristics of Dissolved Oxygen in Typical Water Bodies of South China Based on Landsat Imagery (2013-2024), EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-4512, https://doi.org/10.5194/egusphere-egu25-4512, 2025.
Freshwater ecosystems can experience shifts in aquatic primary production that are driven by disturbances in temperature, river discharge, and nutrient cycle dynamics – with interacting and cumulative effects that are poorly understood, but have strong implications for ecosystem health. These shifts can reduce biodiversity, alter food web structures, degrade water quality, and negatively impact aquatic communities. Changes in the timing of phytoplankton growth cycles are often associated with different algae groups dominating the assemblage. While cyanobacterial blooms are of primary concern due to their toxicity and tight association with high summer temperatures and nutrient loads, other harmful and potentially toxic algae may proliferate as well. In the summer of 2022, the Oder River experienced a harmful algal bloom caused by the brackish-water haptophyte Prymnesium parvum. This unprecedented event culminated in an environmental disaster in one of Europe's last rivers with a free-flowing lower course and several regions of extraordinary ecological importance. Here, we integrate long-term records of hydrometeorological variables with the estimations of chlorophyll content derived from the Copernicus satellite Sentinel-2 along the full extent of the Oder River and its most relevant tributaries to (1) characterize the timing (i.e., phenology of each growth cycle) and magnitude (i.e., peak chlorophyll-a activity) of the phytoplankton dynamics over the last 10 years; (2) evaluate the role of temperature and discharge anomalies, and of increased saline inputs in driving these blooms; and (3) compare these assemblages with those observed in other European and global river systems. Our findings reveal distinct patterns of spring and summer blooms alternating between years in magnitude as well as in onset timing. When incorporating cumulative anomalies of temperature and discharge the sensitivity of the phytoplankton community dynamics to the interplay of environmental drivers becomes clearer. We discuss the implications of these patterns in a context of rapid global change.
How to cite: Torre Zaffaroni, P., Stelzer, K., Scholze, J., Bremerich, V., Lebreton, C., and Goldhammer, T.: Environmental drivers of algal bloom timing and magnitude estimated from Sentinel-2 imagery, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16618, https://doi.org/10.5194/egusphere-egu25-16618, 2025.
Algal blooms, resulting from rapidly growing algae in freshwater and marine environments, pose serious risks to biodiversity, ecosystems, and human health. Algal blooms are predicted to increase due to temperature, nutrient availability, and alien species invasions. Satellite remote sensing products provide a versatile monitoring tool that complements in-situ sampling. This study used remote sensing products to investigate algal bloom dynamics across inland water bodies in England (Lake Windermere) and India (Yamuna and Ganga rivers). Specifically, Google Earth Engine was used to characterise BGA, MCI, and NDCI from high-resolution satellite data to investigate algal blooms in Windermere from 2020 to 2025 and Yamuna and Ganga rivers in 2021 to 2024. By integrating data from Sentinel-2 and PlanetScope, enhanced by UAV sensor technology for high-resolution data collection, we establish predictive models for assessing water quality parameters. To analyze the data, we implement ML algorithms. Our findings indicate that RF outperforms other ML algorithms when using Sentinel-2 data, achieving an overall accuracy of 70.71% with a Kappa statistic of 0.79. Integrating a similar methodology on PlanetScope and high-resolution drone imagery to improve and increase performance boost is an ongoing task. To assess phytoplankton blooms using satellite data, we are analyzing imagery from sources like Landsat-8, 9, MODIS, and Sentinel-2 to quantify the number of blooms based on chlorophyll-A concentrations. The effectiveness of this monitoring depends on the spatial resolution, which influences the detection of smaller blooms (high-resolution imagery captures more detail), and the temporal resolution, which affects the ability to monitor ephemeral events (daily data is optimal) and thus to actually quantify is a challenge per se. Even with the performance metrics, we establish correlations between band indices and in-situ field-based measurements (pH, temperature, salinity, conductivity, turbidity, etc.). An online dashboard application will be developed to visualize results through spectral band wavelength charts, time-series data, and spatial distribution maps by integrating UK and India’s environmental agency open-source data. The future scope of our methodology can incorporate advanced techniques such as SAM, spectral feature fitting, and continuum band removal for quantitative hyperspectral data analysis. This comparative analysis emphasizes the urgent need for continuous monitoring to protect ecosystems and public health in both regions. We advocate innovative, sustainable water resource management approaches by uniting advanced remote sensing technologies with traditional methods. Ultimately, our findings aim to inform interventions to improve water quality and ecological health, benefiting local communities and the ecosystems they depend on. Through collaborative efforts, this study aspires to enhance understanding of the intricate connections between water quality dynamics, paving the way for policymakers to adhere to comprehensive management strategies that address the needs of future generations by focusing on SDG-6 (clean water sanitation) and SDG-14 (life below water).
How to cite: Galodha, A., Peppa, M.-V., Wilson, S., Anees, S., Lall, B., and Ahammad, S. Z.: Dynamic monitoring and mapping of water quality indicators using multi-modal and multi-scale satellite imagery, UAVs, and open-source cloud computing platform, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-19332, https://doi.org/10.5194/egusphere-egu25-19332, 2025.
The trophic index is one of the most important indicators for primary production and potential anthropogenic eutrophication of lakes. In Germany, It is calculated from measured phosphorus concentration, visibility and chlorophyll-a content in water samples collected during the productive period between April and October. These parameters are monitored for most lakes > 50 ha, which are covered by the Water Framework Directive. Monitoring typically occurs only at a low interval of several years, making it difficult to distinguish trends in lake water quality from natural annual variations. Moreover, no information is available for small lakes < 50 ha, thus excluding a high proportion of lakes from trophic state monitoring.
In the presented work, we investigated the extent to which satellite data are able to fill these gaps. There are many indices for real-time water monitoring based on satellite images from the Copernicus Sentinel-2 program. Based on this existing know-how, the reliability of satellite-based trophic index assessment was validated along the following questions:
- which bands of the Sentinel-2 images are best suited for estimating trophic state?
- how does the data need to be temporally aggregated within a season?
- is one pixel of a lake sufficient to reliably describe the trophic state of a lake, so small lakes can be included in the assessment?
The investigation was based on 294 lakes in Brandenburg, Germany. Monitored trophic index from the years 2018 to 2022 was correlated with satellite information for one pixel, chosen randomly in the center of each lake. The trophic index based on in-situ measurements is best calculated from monthly values. Similarly, satellite-derived indices were first averaged monthly and then seasonally (April to October in Germany).
Results show that Bands 2 and 5 of the Copernicus Sentinel-2 Mission are best suited to describe the differences of trophic state. The developed Normalized Difference Trophic Index (NDTrI) is based on single image indeces which are defined as:
Band 5 describes the near infrared reflectance at 705 nm, band 2 the reflectance of blue light at 490 nm. In oligotrophic lakes, band 2 reflectance usually dominates and the index is below zero. The resulting NDTrI was found to be highly correlated with the in-situ data for the available years (Pearson correlation coefficient per year between 0.83 and 0.92). The data are available at an annual resolution, which is three times more frequent than the conventional analysis. This allows a much more reliable trend analysis, which can be used to monitor the success or lack of water quality management more quickly. The feasibility of the methodology, using only one pixel for each lake, indicates that thousands of small lakes can be included in the remote monitoring without much effort.
A first sensitivity analysis has shown that the classification is more reliable for eutrophic water bodies than for oligotrophic ones. Further factors influencing the accuracy of the method and potentials of trend as well as seasonal analysis will be investigated in the European Horizon projects AD4GD and ProCleanLakes.
How to cite: Zamzow, M., Matzinger, A., Rustler, M., and Bastin, L.: Satellite-derived trophic index to support management of small and medium-sized lakes , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-20482, https://doi.org/10.5194/egusphere-egu25-20482, 2025.
River management, e.g. in the context of the European Water Framework Directive, requires a comprehensive monitoring of inland water bodies. Traditional water quality assessment methods using probes and laboratory analyses are time-consuming, expensive and insufficiently capture the large-scale dynamic nature of river systems. Remote sensing offers a promising additional data source, though limited river widths are challenging to resolve and constrain sensor selection and often necessitates integrating multiple data sources of variable resolution. In addition, using sensors with higher resolution enables the investigation of small-scale effects.
The MeskalMon-Project (Multi-scale monitoring in rivers using remote sensing and in-situ methods for the parameters chlorophyll and suspended matter) develops an innovative approach that combines in-situ measurements with remote sensing data from various platforms, including a hyperspectral sensor mounted to bridges, a spectrometer, a multispectral UAS-sensor and multispectral satellite imagery. The research focusses on the characterization of the spatial variability and spectral interactions of chlorophyll-a and turbidity through a comprehensive monitoring strategy.
Measurement campaigns on the river Moselle during 2022-2024 employed diverse sampling techniques, including longitudinal, lateral, and vertical measurements in the water column. By applying indices such as the Normalized Difference Chlorophyll Index (NDCI), we have facilitated comparability between different spatial resolutions, data acquisition methods and platforms. Preliminary results show a promising agreement between satellite, camera, spectrometer and in-situ measurement methods.
Our findings indicate that water, sediments and nutrients in the river Moselle are well mixed, which makes surface data from remote sensing representative despite its limited penetration depth. However, the variable composition of different algae groups in the water and surface scum formation in the case of intensive cyanobacterial blooms, pose major challenges in the interpretation of remote sensing data to derive the concentration of suspended sediment and chlorophyll-a (Chl-a) in the water column.
Analysis of multispectral satellite data (here Sentinel-2) shows good results for Chl-a and turbidity quantification. We achieve high determination coefficients of up to 0.79, using different atmospheric corrections in combination with various algorithms for deriving Chl-a from satellite data using in-situ measurements. Limitations arise if algae groups vary or high Chl-a concentrations are accompanied by high turbidity. The analyses demonstrated the intricate optical interactions within aquatic environments, highlighting the challenges of accurately distinguishing and measuring water quality indicators through remote sensing techniques, showing advantages of hyperspectral methods.
Our research revealed significant variations in the performance of Chl-a algorithms and indices depending on the mix of algal groups present in the water. The current spectral bands available on the used UAS-sensor and the satellites proved insufficient for differentiating algal groups. However, the upcoming CHIME mission provides new opportunities for a more detailed analysis of aquatic ecosystems in the necessary spatial, spectral and temporal resolution and demonstrates the potential of advanced remote sensing technologies.
This research provides a novel, integrated framework for remote sensing based water quality monitoring that overcomes some limitations of traditional monitoring methods. It represents a significant step towards more dynamic, comprehensive, and efficient environmental monitoring strategies, with future research poised to leverage emerging satellite technologies for more nuanced ecological insights.
How to cite: Baschek, B., Brehm, T., Herrmann, M., Koch, D., Klotz, F., Kleinteich, J., Nicholls, C., and Hoffmann, T.: Multi-scale Monitoring of Water Quality in a Phytoplankton Carrying, European River - a case study of the Moselle, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-17411, https://doi.org/10.5194/egusphere-egu25-17411, 2025.
River transport of reactive nitrogen from land to sea is an important component of the nitrogen cycle, significantly influencing freshwater and coastal water quality and their ecosystem health. Human activities have markedly accelerated the Earth’s nitrogen cycle since pre-industrial times. Despite estimates of river total nitrogen export, the speciation of river nitrogen export to global coastal waters and its spatiotemporal changes remain poorly understood. Assessing the long-term changes in the river export of different nitrogen forms to coastal systems worldwide in response to their different trajectories of human perturbations is crucial for developing effective mitigation strategies for nitrogen pollution and improving water quality. In this study, we quantify the river export of different nitrogen forms to global coastal waters from 1900 till 2010 using the spatially explicit, mechanistic, coupled hydrology and biogeochemistry model IMAGE-DGNM. This model keeps track of nutrient supply from the land, perturbations of river network, and hydroclimate change, and describes the dynamic biogeochemical nitrogen transformations and transport along the terrestrial-freshwater continuum. Results show that although the river export of all major nitrogen forms increased during 1900-2010 at the global scale, some regions have shown stable or decreasing trends in recent decades. Moreover, the composition of different forms in river nitrogen export differs across different regions. Not only the fluxes but also the fractions of different forms changed differently across systems during 1900-2010, which emphasizes the importance of taking into account varied human impacts, climates and hydrological conditions to address the complexity of mitigating local nitrogen pollution.
How to cite: Wang, J., Liu, X., Vilmin, L., Beusen, A. H. W., Bouwman, A. F., and Middelburg, J. J.: Divergent changes in river nitrogen export to coastal waters worldwide in the Anthropocene, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-20321, https://doi.org/10.5194/egusphere-egu25-20321, 2025.
River basins have been extensively altered by unsustainable surge inagriculture, industrialization and urbanization. Green remediation, particularly phytoremediation, is an eco-friendly and efficient approach that utilizes plants and their associated microbes to remove, detoxify, or immobilize toxins from soil, water, or air, thereby enhancing water quality by addressing contaminants such as metals and nutrients. A lab-scale wetland study was carried out to determine the efficacy of a specific plant species for phytoremediation. This study analyzed pollutant parameters, including metals (As, Al, Ca, Cd, Co, Cu,Fe, Pb, Mg, Hg, Ni, Na) and nutrients like nitrogen (nitrates and ammoniacalnitrogen) and phosphorus (total phosphorus and available phosphorus). CannaIndica, an aquatic macrophyte, was used for its nutrient and metal removal capabilities. A wetland simulator, fabricated from acrylic sheets (length: ~1m,height: ~0.75m, width: ~0.50m), was used to grow Canna Indicaplants. Thewetland simulator was filled with a 150 L volume (40% of reactor volume) ofgrowing medium (soil, sand, gravel), arranged in a block design. A water samplefrom Ratanpuri in the Hindon River, known for its high pollution levels in northern India, was used for the wetland study, which was analyzed for metalsand nutrients before, during, and after the experiments. The experiments were performed for 20 days (three runs), depending upon their treatment efficiency.This study's findings demonstrated that metals' removal efficiency is 50-55 percent (absorbed by plants). Similarly, the efficacy of nutrient removal,specifically nitrogen and phosphorus compounds, using phytoremediation is evaluated in this study, with removal rates ranging from 60-74 percent. The findings highlight phytoremediation's performance as a highly sustainabletechnology for remediating contaminated water bodies or soil structures. As urbanization and industrialization accelerate, rising river contamination levels have increased the lateral flow of pollutants from riverbanks into groundwater. So, implementing field-scale green remediation strategies using Canna Indica plants along riverbanks mitigates contaminant movement, ensuring soil and water quality restoration amidst rising anthropogenic demands.
How to cite: Bodana, D., N Srivastava, A., Vinnarasi, R., and Kumar Jain, S.: Green remediation of River Water Contaminants: A Lab-Scale Wetland Approach for Metal and Excess Nutrient Removal, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-1113, https://doi.org/10.5194/egusphere-egu25-1113, 2025.
Changing weather patterns and extreme hydrological events, i.e. heavy rainfall and prolonged droughts, have resulted in further degradation of water quality within the agricultural landscapes by exacerbating nutrient transfer processes to surface water bodies. More frequently occurring extreme weather events require development of robust adaptation/mitigation strategies, which further requires an improved understanding of both the timing and conditions of the intensified hydro-meteorological drivers.
In order to evaluate the impact of extreme events on nutrient losses, an empirical modelling approach was taken in six intensively-monitored hydrologically-diverse agricultural catchments (ca 3-30 km2) across Ireland. The objective was to link the occurrence of pulses of N and P, driven by hydrology, with sub-hourly water quality monitoring and weather data over a 14-year period. Then using the downscaled future climate change scenarios for moderate (RCP 4.5) and severe (RCP 8.5) emission pathways, the occurrence probability and timing of such triggering events were investigated for three different time periods until end of the century.
The investigation of high-temporal resolution data confirmed capturing all subtle changes in the nutrient concentrations and extreme weather events while the empirical modelling of associated nutrient losses events due to extreme hydrological events revealed various criteria contributing to nutrient-losses. These criteria were in terms of air temperature and effective rainfall and explained more than 50% of any nutrient loss events across all the catchments at different temporal scales. Temporal aspects of data analysis showed that certain months would require specific attention in terms of adaptation, management, and re-evaluating nutrients’ pathways.
Comparison between RCP 4.5 and 8.5 across three time periods of near future (2010-2039), mid-future (2040-2069), and far-future (2070-2100), suggested that the upward trends in number of events continue to increase stepwise in each time period whereas the percentage increase of nutrient-concentrations’ increasing events would almost double in RCP8.5. There would be over 60% and 40% increase in the number of P-loss and N-loss triggering events, respectively, from near-future to far-future considering the sum of different empirically-driven criteria. Meanwhile, the catchment characteristics played a major role in defining the response of each landscapes to various drivers. Such catchment-specific response was explained by hydrological connectivity, soil chemistry and texture, drainage status, and agricultural practices.
Prolonged drought and warm periods and increased hydrological connectivity would result in increasing number of nutrient losses events as we move toward end of the century. The detected differences in catchments’ characteristics and in the frequency of triggering events across climate scenarios were indicative of consequence for future mitigation strategies and policy decisions which have to be climate smart, resilient, catchment-specific, and tailored to different environments
This research has been conducted as part of the WaterFutures project (Irish-EPA-funded).
How to cite: Ezzati, G. and Mellander, P.: An empirical modelling approach to investigate nutrient losses in view of more frequent extreme hydrological events using future climate-scenarios, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-3515, https://doi.org/10.5194/egusphere-egu25-3515, 2025.
Catchment-scale nitrate dynamics involve complex interactions coupling hydrological transport and biogeochemical transformations, imposing challenges for source control of diffuse pollution. Here, we propose a novel spatio-temporal framework for catchment-scale quantification of Damköhler number (Da) based on the ecohydrological modelling platform EcH2O-iso and a catchment nitrate module. We examined Da variability of the dominant process of denitrification and N removal in the intensively instrumented, heterogeneous Selke catchment (456 km2, central Germany). Results showed that warm-season N losses from denitrification was of catchment-wide significance (Da >1), while its high spatial variations were co-determined by varying exposure times (e.g., hydrologically isolated areas with long residence times and old water) and removal efficiencies (e.g., hotspots of channel-connected lowland areas). Moreover, Da demonstrated a systematic shift to transport-dominance during the wet-spring season (from >1 to <1). Under the prolonged 2018-2019 droughts, denitrification removal generally reduced, resulting in further N accumulation in agricultural soils. Besides, the hydrologically disconnected lowland areas (with high water ages) exhibited extra risks of groundwater contamination. Importantly, the channel-connected lowlands exhibited high removal efficiencies, as well as high resilience to the disturbances like the droughts. These insights into integrated catchment functioning highlighted important management implications for nature-based, spatially targeted mitigation measures.
How to cite: Yang, X., Tetzlaff, D., Jin, J., Li, Q., Borchardt, D., and Soulsby, C.: A Damköhler-based catchment nitrate transport-processing integration and its responses to droughts, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-14220, https://doi.org/10.5194/egusphere-egu25-14220, 2025.
The European Water Framework Directive (WFD) 2000/60/EC is an important milestone for water management, aiming to achieve good chemical and ecological status for European water bodies. Achieving these goals requires a comprehensive understanding of nutrients and pollutants transfer through surface waters, especially in areas impacted by human activities. This study aims to develop a methodological framework for setting up numerical models at the Northeast Italian River district scale to analyze nutrients and pollutants transfer from agriculture, industrial production, and wastewater treatment. The research enhances knowledge on the distribution of priority substances, supporting effective water management strategies aligned with WFD objectives.
The study area covers about 12,000 km² across 11 river basins in Veneto and Friuli Venezia Giulia. Separate models were developed for each basin using the SWAT (Soil and Water Assessment Tool) model. The models focused on simulating river discharge, calibrated against observed data using the Nash-Sutcliffe Efficiency (NSE) coefficient, and estimating nutrient and phytochemical loads discharged to the sea, verified with literature and limited monitored data. Given the highly modified river network, simplifications were introduced. Irrigation diversions and channels were modeled as point sources, along with industrial and wastewater treatment discharges. To account for agricultural practices, a detailed land cover layer was created by integrating Corine Land Cover 2018 and EUCROP 2018 datasets. Fertilizer and phytochemical use were defined through specific scheduling for all major crops, covering up to 95% of the study area. Current climate conditions were simulated using observed data from 2001 to 2020. Future scenarios for Global Warming Levels of 2°C and 3°C were modeled using downscaled data from Regional Climate Models.
Results show a good reproduction of discharge rates across river basins, with monthly NSE of 0.5 or higher values. NSE values below 0.5 were observed in the Venice Lagoon basin, the most anthropized area, where tidal effects were not captured by SWAT. Nutrient loads of total nitrogen and total phosphorus discharged to the sea aligned with previous studies, highlighting key point and area sources within each basin. Under future climate scenarios, both annual flow and nutrient discharges are expected to slightly increase. The study results were used by the Northeast Italian River District Authority to develop recommendations for the improvement of surface water monitoring and management strategies, so to contribute to an increasingly effective implementation of the WFD goals.
How to cite: Camera, C. A. S., Pedretti, D., Dalla Libera, N., Pasini, S., Gelmini, Y., and Braidot, A.: Methodological framework for the analysis of current and future nutrient and pollutant fluxes in the lowlands of Northeast Italy, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16080, https://doi.org/10.5194/egusphere-egu25-16080, 2025.
The Soil and Water Assessment Tool (SWAT) is a widely used hydrological model designed to simulate water flow, sediment transport, and nutrient cycling in complex river basins. This study employs SWAT to assess the impact of both diffuse and point sources of nutrient contamination on river water quality in the Adige River basin, one of Italy's largest and most significant waterways. The Adige River basin presents a diverse hydrological system influenced by agricultural practices, urban development and industrial activities, making it a representative case for evaluating nutrient dynamics and their environmental implications in a complex contest.
Diffuse sources of nutrient pollution, primarily from agricultural runoff, contribute substantially to the nutrient load in the Adige River, particularly nitrogen and phosphorus. These nutrients often originate from fertilizers, animal manure and soil erosion and their impacts are exacerbated by rainfall and irrigation practices. Point sources, such as wastewater treatment plants, industrial effluents, and urban discharge points, introduce localized but often concentrated nutrient loads to the river system. Understanding the interaction between these sources is critical for developing effective management strategies which prioritize interventions to mitigate adverse effects on water quality.
Using SWAT, this study integrates extensive spatial and temporal data, including land use, soil properties, climate variables and hydrological records, to simulate nutrient fate and transport across the basin. The model’s ability to account for both diffuse and point sources allows for a holistic understanding of nutrient dynamics and investigating their cumulative and disaggregated impacts on the river. SWAT outputs are validated against observed water quality data, ensuring robust and reliable simulations.
The application of SWAT in the Adige River basin not only highlights the spatial distribution and seasonal variability of nutrient loads but also identifies critical source areas that disproportionately contribute to contamination. By simulating various land-use and management scenarios, SWAT provides actionable insights into how different interventions, such as buffer strips, reduced fertilizer application, or advanced wastewater treatment, can mitigate nutrient pollution. Furthermore, the model supports the identification of nutrient retention and removal hotspots within the river system, enhancing our understanding of the natural attenuation processes that influence contaminants fate.
The insights gained from SWAT modelling in the Adige River basin have broader implications for water quality management in similar river systems worldwide. The model's comprehensive approach to linking hydrological processes with nutrient dynamics strengthens the knowledge base for addressing diffuse and point source contamination. It offers a scientific foundation for formulating efficient guidelines and policies aimed at reducing nutrient inputs, improving wastewater management, and restoring aquatic ecosystems.
In conclusion, SWAT serves as a powerful tool for understanding the complex interplay between diffuse and point sources of contamination and their effects on river water quality. Its application in the Adige River basin underscores its utility in advancing knowledge about nutrient fate and transport and in guiding targeted, evidence-based strategies to mitigate contamination. This study highlights the pivotal role of hydrological models like SWAT in achieving sustainable river basin management and protecting water resources from the growing pressures of human activities.
How to cite: Dalla Libera, N., Pasini, S., Gelmini, Y., Pedretti, D., and Camera, C.: SWAT modelling to unveil and manage the impact of diffuse and point source pollution within Adige River basin , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-17952, https://doi.org/10.5194/egusphere-egu25-17952, 2025.
Gorgan Bay, situated in the southeastern Caspian Sea, is an ecologically significant wetland and a critical habitat for wildlife, migratory birds, and fisheries. However, the bay's ecological and economic importance is increasingly threatened by eutrophication, driven by substantial nutrient inflows, particularly nitrogen (N) and phosphorus (P), originating from its upstream catchment. This study aims to quantify nutrient loads, identify their sources, and evaluate their impacts on water quality using land-use analysis, export coefficient modeling (ECM), and the Carlson Trophic State Index (CTSI). Over the past two decades, nutrient loads have risen steadily, with phosphorus and nitrogen inputs each increasing by approximately 3.6% due to agricultural intensification, urbanization, and untreated wastewater discharge. Concurrently, the annual discharge of freshwater into Gorgan Bay has shown a significant declining trend, despite minimal changes in monthly flows. The Qarasu River, contributing approximately 50% of the total inflow, and the Baghou River, accounting for 14%, play crucial roles in the bay’s hydrological balance. The declining water inflows, coupled with high evaporation rates and no significant replenishment since 2015, have led to a persistent decrease in the bay’s water storage. Notably, in 2017, the water loss exceeded 2 million cubic meters. These changes have resulted in reduced water quantity, which directly affects water quality and intensifies eutrophication in the bay. Land-use changes have further exacerbated nutrient export. Analysis indicates a 10% increase in agricultural land, often linked to intensive fertilizer use, and a 7% reduction in forested areas, which has diminished the natural capacity for nutrient filtration. The CTSI analysis reveals that while some areas of Gorgan Bay are mesotrophic, most central and eastern regions are classified as eutrophic, reflecting significant seasonal and spatial variations in water quality, algal productivity, and light penetration. These changes have resulted in severe ecological consequences, including algal blooms, reduced water clarity, and hypoxic conditions, which pose substantial threats to aquatic biodiversity and ecosystem stability. Furthermore, the socio-economic ramifications for fisheries and local communities reliant on the bay's resources are profound. To address these challenges, this study recommends adopting integrated management strategies. Key measures include controlling point-source phosphorus pollution through advanced wastewater treatment, promoting sustainable agricultural practices to optimize nitrogen and phosphorus use, and restoring riparian vegetation to enhance natural nutrient buffering. Additionally, increasing environmental awareness and fostering stakeholder engagement at the catchment scale are crucial for achieving long-term conservation goals. Implementing these strategies can help mitigate eutrophication impacts, improve water quality, and preserve the ecological and economic significance of Gorgan Bay for future generations.
How to cite: kazemi, M., Hossaini Baheri, M., Mirzajani, M., Fakhr, M., and Tajrishy, M.: Assessing Eutrophication Drivers and Water Quality Degradation in Gorgan Bay: A Catchment-Scale Nutrient Export Analysis, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-7804, https://doi.org/10.5194/egusphere-egu25-7804, 2025.
Total organic carbon (TOC) in surface water significantly impacts the global carbon cycle, ecosystem productivity, and potable water quality. Although physically based watershed models, such as the Terrestrial-Aquatic Sciences Convergence (TASC) model, can simulate carbon cycling in surface waters, challenges persist in representing groundwater contributions due to limitations in TASC's groundwater module. Additionally, like its foundation, the Soil and Water Assessment Tool (SWAT), the TASC model struggles to accurately represent distributary stream systems. In this study, we modified the TASC model to examine the carbon budget of the Guangdong-Hong Kong-Macao Greater Bay Area (GBA), a densely populated coastal delta of the Pearl River. The modified model successfully simulated the carbon dynamics of the region’s distributary stream system, estimating a dissolved organic carbon (DOC) flux of 1.05 × 10⁹ g/day across eight watershed outlets. TOC in the western outlets was primarily influenced by suspended sediments, while TOC in the eastern outlets originated mainly from agricultural runoff and domestic sewage. Furthermore, seasonal variations revealed important patterns in hydrological contributions. Groundwater flow contributed significantly to river discharge during the winter months (November to January), occasionally surpassing overland runoff. Remarkably, groundwater DOC fluxes dominated riverine DOC throughout the year, accounting for 41% to 62% of the total DOC contribution. This study highlights groundwater as a vital pathway for the transport and release of dissolved carbon into rivers, representing a potentially significant carbon loss within watersheds.
How to cite: Zhan, Y. and Guo, Z.: Groundwater contributions to the carbon budget of the Greater Bay Area, China, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-2453, https://doi.org/10.5194/egusphere-egu25-2453, 2025.
In many regions of the Netherlands, ammonium concentrations in surface water exceed the EU Water Framework Directive standard of 0.3 mg NH4-N/L, which has an ecological origin. In many Dutch polders, there is significant seepage of groundwater that may contain naturally high concentrations of ammonium. This combination leads to a background load that is high compared to anthropogenic loads by agriculture, waste water discharge and other sources. The Flevoland Polders that were established in the 1960s, are known to be such an example.
This study investigated the water quality and surface water load with respect to ammonium for the Flevoland Polders for which large datasets on groundwater quality and surface water quality are available. We aim to characterise : 1. the variation of ammonium concentration in surface water, seasonally and between dry and wet years, 2. the spatial variation of groundwater ammonium concentration for different time intervals, and 3. the contribution of ammonium from groundwater to the surface water system compared to other sources such as waste water treatment plants, drain water from agricultural land, and leaching from nature areas.
The typical concentration ranges for ammonium are as follows: 1. about 0 – 8 mg NH4/L for ditches and canals, 2. from 0 to 60 mg/L or even higher for groundwater, and 3. approximately 0 – 8 mg/L for drain water in agricultural areas. This illustrates the major role of groundwater exfiltration in the ammonium load of surface water taking into account groundwater exfiltration rates compared to recharge by net precipitation. More detailed data analysis indicates seasonal variations in surface water ammonium concentrations, with higher levels in winter, likely due to reduced microbial activity, and lower levels in summer, although occasional summer peaks were observed. No significant patterns were found between wet and dry years, possibly due to the lack of extreme wet or dry conditions during the study period. Further, the water analyses show that groundwater ammonium concentrations were relatively stable over time, except for data prior to 1980. Mass balance calculations indicate that groundwater seepage is the major source of ammonium followed by drain water from agricultural land. Waste water treatment plants and nature areas contributed the least.
Overall, this study highlights the significant role of groundwater in contributing to ammonium loads in surface water. It also shows that background load should be taken into account when establishing water quality standards for surface water.
How to cite: Griffioen, J. and Zhang, H.: The importance of exfiltrating nutrient-rich groundwater to the ammonium load of surface water in the Flevoland Polders, the Netherlands, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-3261, https://doi.org/10.5194/egusphere-egu25-3261, 2025.
The application of stable isotope techniques toward the identification and quantification of nitrogen pollution has been a topic of significant interest in recent decades. The overlap of isotopic signatures for animal manure, septic waste, soil derived nitrogen, and ammonium-based fertilizers can obscure the source of pollutants, which are desirable to identify. To overcome this overlap of signals, the utilization of co-tracers, such as chemical compounds, which can be attributed to specific source inputs, becomes advantageous. An approach, which has commonly been utilized to identify chemical compounds – compounds of emerging concern (CECs) in this context – is the utilization of quantitative structure-activity relationship (QSAR) models. Through QSAR models, CECs may be identified and grouped based on known physical and chemical properties to predict their behavior in different scenarios. To better delineate sources of nitrate pollution, a QSAR analysis was applied with the aim of identifying CECs, which act as conservative environmental tracers and are specifically linked to nitrate pollution sources. For this purpose, stable nitrate isotope data was coupled with CEC data collected from several case study sites in Austria, France, Greece, and Nigeria. Additionally, nitrate isotope data was introduced to a Bayesian mixing model (MixSIAR) to delineate potential pollution sources. The QSAR model results revealed the CECs with high potential, and were compared with those of the MixSIAR model in order to identify those compounds, which could be strong candidates as co-tracers in nitrate pollution studies. Further, the parameters of the QSAR model were extracted to identify compound specific parameters which may be indicators for other compounds with tracer potential in future studies. The goal of this work was to identify CECs or types of CECs as robust co-tracers for nitrate source determination in order to better understand N cycling in ecosystems and contribute towards the protection and sustainable management of water resources.
How to cite: McGuire, B., Harjung, A., Matiatos, I., Re, V., Etuk, M., Huneau, F., and Vystavna, Y.: Advancing Nitrogen Source Identification with Compounds Emerging Concern Co-Tracers and Isotope Analysis in Natural Waters, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-17299, https://doi.org/10.5194/egusphere-egu25-17299, 2025.
Phosphorus (P) remaining in the farming system after accounting for all P inputs (chemical fertilisers, imported slurry/manure, concentrate feed) and offtakes (agricultural products, exported slurry/manure) can be summarised as the farmgate P balance (FPB). If FPB is in surplus, P can be immediately vulnerable to runoff and also to become part of a legacy soil P store that can have longer term diffuse pollution consequences. The FPB scaled to Northern Ireland was investigated against national water quality datasets from one hundred and one rivers over 18-years. National FPBs ranged from 8.7 – 15.1 kg P/ha/yr over that period. Ninety-three of the river sites were used for Water Framework Directive (WFD) reporting of baseline soluble reactive P (SRP) concentrations. Total P (TP) data from eight major rivers were combined with river discharge for large catchment area (4,836 km2) estimates of TP load to the 304 km2 Lough Neagh lake basin where lake TP is a WFD reporting requirement. River TP loads were normalised to annual flow-weighted mean concentrations (FWMC). Based on conceptual models of ‘catchment memory’ and rivers as ‘jerky conveyor belts’ of material movement, the study found a linear 1-year lag between annual FPB and mean annual baseline SRP concentration in the ninety-three river sites, and a 5-year lag between FPB and TP FWMC in the eight major rivers. The differences were explained by soluble and particulate P partitioning and fate between source and receptors. The linear model suggested that river SRP would need a stronger FPB to aim for the good/high SRP boundary (upper quartile 0.037 mg/L) to meet a river FWMC TP concentration (0.109 mg/L) that would approach the lake’s WFD TP moderate/good boundary (0.044 mg/L) in the absence of internal lake P loading. The analysis suggested that the FPB would need to be 5.5 kg/ha/yr, or half the current balance, to meet these targets in the absence of other non-FPB sources. Addressing these non-FPB sources of P pollution in the Lough Neagh catchment and Northern Ireland more generally (38% considered to be from urban and rural sewered populations), would either speed up and attain higher water quality targets, or relieve the burden to agriculture if a slightly higher FPB target was used.
How to cite: Jordan, P., McElarney, Y., and Cassidy, R.: Using the farmgate phosphorus balance to meet river and lake water quality targets , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16790, https://doi.org/10.5194/egusphere-egu25-16790, 2025.
The analysis of nutrient dynamics during precipitation events using normalised cumulative loads (NCL) is a classic and frequently used method of water quality analysis. Normalised pollutographs combine the cumulative nutrient load and cumulative runoff of an event in one plot to better interpret the nutrient export dynamics of a catchment. Essentially, the strength of a first-flush event or an ongoing dilution can be quantified using NCL. However, different runoff generation processes may be superimposed, hampering process analysis. Thus, our approach combines the classical pollutograph with hydrograph separation using stable water isotopes. The pollutograph is set up for each of the separated runoff components (pre-event water and event water) and the export dynamics of the runoff components of the same event are compared with the dynamics of the total runoff via the calculated area under the curve. K-Means cluster analyses were used both to categorise the functional behaviour of the runoff components and to identify any changes in export dynamics between the runoff components.
This method was applied in the catchment area of the ‘Nesselbach’, which is located near Hofgeismar (district of Kassel) in the northern foothills of the West Hessian depression. The catchment area is primarily characterised by agricultural land use, which is accompanied by a small forest and a small settlement. All necessary climate parameters were recorded by a weather station located in the catchment. The sampling period extended from 1 February 2021 to 1 August 2022. Measurements of nitrate concentrations and runoff were recorded using high-resolution optical in situ probes (resolution: 5 min), supported by automatic samplers to collect phosphorus, major ions and isotope samples during 15 precipitation events (resolution: 15 min). All parameters were additionally calibrated by regular manual sampling with laboratory values.
Using the extended method, our results show clear differences in the export dynamics of phosphorus and nitrate, as well as unexpected results in the direct comparison of export dynamics between total runoff and event water. Phosphorus shows similar dynamics in the event water as in the total runoff (of the same event), with a tendency to more extreme values. For nitrate, on the other hand, the dynamics of nitrate export sometimes changed drastically; export dynamics tending towards dilution in the total runoff show wash-off tendencies in the event water and vice versa. The same approach was also used to investigate the export dynamics of the ions Na, Ca, SO4, K, Mg and Cl, which served to increase the sample size of the analyses carried out and thus to confirm the robustness of the method.
How to cite: Ditzel, L., Spill, C., and Gaßmann, M.: A methodological extension of the classic normalised cumulative loads analysis: combination of normalised cumulative loads and hydrograph separation for nutrient exports during precipitation events., EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-2531, https://doi.org/10.5194/egusphere-egu25-2531, 2025.
Nutrient fluxes exhibit significant spatiotemporal heterogeneity, driven by the dynamic coupling of hydrological processes and biogeochemical cycles. To elucidate the mechanisms controlling nutrient species, we conducted a process-based investigation of the Pearl River Basin using simulations from the Soil and Water Assessment Tool. Our findings reveal that surface nitrate and soluble phosphorus are primarily transported by surface flow, whereas particulate inorganic phosphorus is predominantly regulated by sediment transport. These transport mediators, coupled with soil nutrient pools and fertilizer use, govern the fluxes of nutrient species. Fertilizer use, despite being less abundant than soil pools, exerts a stronger influence due to its closer coupling with transport mediators. In contrast, lateral nitrate flux is largely controlled by soil nitrate pools, with smaller contributions from the coupling between lateral flow and fertilizer use. Organic nutrient species, on the other hand, are primarily regulated by sediment transport in conjunction with plant residues and show minimal dependence on fertilizer inputs. Furthermore, the dominant influence of transport mediators, such as flow and sediment, results in a pronounced wet-season dominance in annual nutrient loads across all nutrient species. Land surface processes further shape the spatial patterns of nutrient fluxes. Surface and lateral flows are most active in regions with high precipitation, with surface flow dominating in urban and agricultural areas, while lateral flow is more prominent in clay-rich forests. Sediment yield is highest in clay-rich urban and agricultural landscapes. Soil nitrate pools and plant residues are abundant in forested regions, whereas inorganic phosphorus pools are elevated in pasturelands. Overall, this study provides critical insights into nutrient dynamics within heterogeneous watersheds, highlighting the interplay between transport processes and nutrient sources across diverse terrestrial landscapes.
How to cite: Zhang, Y. and Gan, J.: Nutrient fluxes and speciation shaped by source-hydrology coupling in the Pearl River Basin, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-4700, https://doi.org/10.5194/egusphere-egu25-4700, 2025.
Over the past seven decades, Germany has undergone transformative changes in wastewater management, largely driven by technological advancements, policy interventions, and the introduction of European Union (EU) directives targeting wastewater treatment plants (WWTPs). Simultaneously, the country has experienced profound societal transformations, notably the political and economic divergence between East and West Germany and shifts in population density, which further influenced WWTP infrastructure and management practices. This study focuses on nitrogen in effluent from WWTPs, which directly discharge into rivers, often having an immediate and localized impact. Understanding the spatial and temporal evolution of nitrogen in wastewater effluent contribution to stream water quality deterioration is essential for designing sustainable water management strategies. To this end, we combined data-driven analysis and modeling approaches, making use of recently published datasets on diffuse nitrogen sources (Batool et al., 2022), nitrogen point sources (Sarrazin et al., 2024), and a state-of-the-art water quality model (Nguyen et al., 2022). We applied the model across various German catchments with diverse agriculture and wastewater amount and treatment development from 1950 to 2020. Our results reveal a noticeable decrease in N effluents from WWTPs, leading to a decline in N contribution to instream nitrogen in the last decades. However, this declining pattern and trend varied across West and East Germany. Our study enables the identification of hot spots, helping spatially targeted management.
References
Batool et al., (2022). https://doi.org/10.1038/s41597-022-01693-9
Nguyen et al. (2022). https://doi.org/10.1029/2022GL100278
Sarrazin, et al. (2024). https://doi.org/10.5194/essd-16-4673-2024, 2024
How to cite: Nguyen, T., Musolff, A., Ebeling, P., Sarrazin, F., Fleckenstein, J., and Kumar, R.: To what extent does nitrogen in wastewater effluent contribute to stream water quality deterioration in Germany?, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-5653, https://doi.org/10.5194/egusphere-egu25-5653, 2025.
The PERUN project (SS02030040) is evaluating water quality data in selected watercourses with major changes in long-term hydrological characteristics. For the purpose of this paper, time series of basic physicochemical parameters will be evaluated, which at some monitoring sites start as early as the 1960s. The evolution of concentrations, their relationship to stream flows, and the objectives of good ecological status given by the Water Framework Directive will be assessed. Problem areas with significant discharges of wastewater and the occurrence of dry periods will be characterised, and the statistical significance of trends in concentrations of physico-chemical parameters will be tested at selected sites with long time series. The typical behaviour of these chemical substances will be described.
How to cite: Barešová, L., Kodeš, V., and Roztočilová, H.: Water quality in watercourses with long-term changes in the hydrological regime, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-5948, https://doi.org/10.5194/egusphere-egu25-5948, 2025.
The present study aimed to analyze to what extent nearby agricultural and semi-natural land use types might affect stream turbidity and chemistry over time. A four-year-long (2021-2024) data on water turbidity and chemistry were measured at different points of the small stream, while soil water content (SWC), and soil temperature were measured at a nearby cropland site with crop rotation.
We analyzed water samples collected daily from the same collection point, and bi-weekly to monthly along the stream from the spring to the outlet (7 to 10 measurement points), whenever water flow was present. We measured stream water turbidity (FNU), total dissolved inorganic nitrogen (as NO3+NO2 and NH4; TDIN) content, water pH, and specific conductance (SPS) using a ProDSS YSI Instrument. Total nitrogen (Ntot) and total phosphorus (Ptot) concentrations were measured using a Nanocolor VIS-II spectrophotometer (Macherey-Nagel) in 2024. Meteorological data was collected from the catchment outlet, while several rain gauges (ECRN-100, Decagon Devices) were also placed at different parts of the small catchment. SWC measurements were collected using 5TM sensors (Decagon Devices) at 10-minute intervals at 15 cm depth.
During the first three years, our results showed a weak correlation between FNU and precipitation (r=0.16, p=0.21), due to high FNU values from low water levels. This mainly occurred during drought conditions. Weak negative connections were observed between SPS and FNU values (r=-0.17, p=0.18), showing that high precipitation lowers water conductivities. Our results showed that the SPC values were inversely proportional to the FNU values.
Based on the 2024 data (n=266) we noted that Ptot levels varied among sites; however, not significantly (p>0.05). Ntot contents were the highest at the site with fresh spring water entering the stream and the lowest at the outlet (p>0.05). We found the strongest correlation to water FNU with orthophosphate (r=0.92; p=0.0003), while a strong correlation between stream discharge and orthophosphate (r=0.65; p=0.005) or total phosphorus concentrations (r=0.63; p=0.006) were also noted. Also, precedent soil moisture content weakly but significantly affected water stream turbidity (r=0.25; p=0.007).
We ran a cluster analysis to determine different levels of rainfall amounts causing significant changes in turbidity values. Three main clusters were distinguished based on the daily sample data, which divided our dataset into daily precipitation amounts of i) precipitation sums below 4.8 mm, ii) averaging 6.3 mm, and iii) averaging 23.7 mm. The three clusters, especially the extreme events are most significantly separated along precipitation and FNU values.
Acknowledgments: This material is based upon work supported by the Hungarian National Research Fund (OTKA/NKFI) project OTKA FK-131792. The research presented in the article was carried out within the framework of the Széchenyi Plan Plus program with the support of the RRF 2.3.1 21 2022 00008 project.This project has received funding from the European Union’s Horizon 2020 research and innovation program under grant agreement No. 862756 (OPTAIN). The research was funded by the Sustainable Development and Technologies National Programme of the Hungarian Academy of Sciences (FFT NP FTA).
How to cite: Horel, Á., Farkas, C., Bódi, A., Zagyva, I., and Zsigmond, T.: Spatiotemporal variations in stream chemistry and suspended sediment levels , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-8457, https://doi.org/10.5194/egusphere-egu25-8457, 2025.
Terrestrial carbon is a crucial source of streamflow dissolved organic carbon (DOC) in catchments. Within river catchments, land use types strongly affect the soil physical and chemical properties, shaping DOC formation, storage, and transport processes. Understanding the relationship between land use types and DOC dynamics is essential for predicting carbon fluxes and mitigating the adverse effects of land use changes on aquatic environments. This study aims to improve the Hydrological Predictions for the Environment (HYPE) model to quantify DOC contributions from different land use. The current HYPE model does not distinct land use type in some DOC processes (e.g., runoff delay.) Our work is to improve the DOC module in HYPE to investigate the land use effects on DOC-related processes from terrestrial to aquatic ecosystems. The DOC processes in each land use type will be characterized by a unique parameter set, which will account for variations in soil organic carbon content, microbial activity, and hydrological transport processes. This approach enables the HYPE model to capture the unique DOC dynamics associated with different land use types. The improved model will be applied and evaluated in the boreal Krycklan catchment in northern Sweden, a region dominated by forests but also including other land use types, such as wetlands and agricultural lands. We aim to answer the research question: How do different land use types influence DOC concentrations in streamflow within boreal catchments? Characterizing the spatiotemporal patterns of DOC contributions from each land use type will provide new insights into the interactions between the terrestrial and aquatic carbon cycling. Additionally, scenarios modeling will allow us to more reliably predict how future changes in land use may affect DOC concentrations and water quality in boreal catchments. Insights derived from this study will provide decision support for sustainable land and water resources management to mitigate the adverse effects of land use changes on aquatic ecosystems.
How to cite: Guo, R., Popp, A., Berggren, M., Liu, J., Liu, J., and Duan, Z.: Modeling the effects of land use on dissolved organic carbon in boreal catchments using the HYPE model, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-12329, https://doi.org/10.5194/egusphere-egu25-12329, 2025.
Quantifying phosphorus (P) loads from watersheds at a fine scale is crucial for studying P sources in lake or river ecosystems; however, it is particularly challenging for mountain–lowland mixed watersheds. To address this challenge, we proposed a framework to estimate the P load at the grid scale and assessed its risk to surrounding rivers in a typical mountain–lowland mixed watershed (Huxi Region in Lake Taihu Basin, China). The framework coupled three models: the Phosphorus Dynamic model for lowland Polder systems (PDP), the Soil and Water Assessment Tool (SWAT), and the Export Coefficient Model (ECM). The coupled model performed satisfactory for both hydrological and water quality variables (Nash–Sutcliffe efficiency > 0.5). Our modelling practice revealed that polder, non-polder, and mountainous areas had P loads of 211.4, 437.2, and 149.9 t yr-1, respectively. P load intensity in lowlands and mountains was 1.75 and 0.60 kg ha-1 yr-1, respectively. A higher P load intensity (> 3 kg ha-1 yr-1) was mainly observed in the non-polder area. In lowland areas, irrigated cropland, aquaculture ponds and impervious surfaces contributed 36.7, 24.8, and 25.8% of the P load, respectively. In mountainous areas, irrigated croplands, aquaculture ponds, and impervious surfaces contributed 28.6, 27.0, and 16.4% of the P load, respectively. Rivers with relatively high P load risks were mainly observed around big cities during rice season, owing to a large contribution of P loads from the non-point source pollution of urban and agricultural activities. This study demonstrated a raster-based estimation of watershed P loads and their impacts on surrounding rivers using coupled process-based models. It would be useful to identify the hotspots and hot moments of P loads at the grid scale.
How to cite: Li, Q., Huang, J., Zhang, J., and Gao, J.: A raster-based estimation of watershed phosphorus load and its impacts on surrounding rivers based on process-based modeling, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-14155, https://doi.org/10.5194/egusphere-egu25-14155, 2025.
The Bay of Brest (BB) is a macro-tidal estuarine environment that has been exposed to strong anthropogenic pressures over the last decades, especially after the Second World War. It is therefore considered as a regional pilot site for addressing coastal ecosystem transformations since the Industrial Revolution. We analysed 4 sediment cores collected in 2 different BB areas more or less exposed to marine hydrodynamic processes: i) Elorn sector (3 cores) and ii) Bay of Daoulas (1 core), in the inner BB, close to the mouth of the Daoulas river, with the aim of deciphering past environmental changes at a high temporal resolution (sub-decadal) over the last 150 years.
Working at a local spatial scale (BB) allows addressing robust correlations between driving forces and environmental changes, as previously demonstrated by pluridisciplinary approaches in the study area (Lambert et al., 2018; Siano et al., 2021). In this project, we are therefore building on this existing dataset (low resolution palynology on the Daoulas core and sedaDNA analyses in the 4 study cores) with the addition of new analyses (high resolution palynology in the Daoulas core and new data for the others, benthic foraminiferal assemblages, sedimentological data and ICP-AES elemental geochemistry in the Daoulas core), with a very fine study resolution (2-year resolution for the Daoulas and 2 to 16-year resolution for the others). Furthermore, statistical analyses based on paleoecological time series allow the detection of major break points, especially thanks to the palynological and sedimentological datasets.
This work then highlights 3 major thresholds (1945, 1960-170, 1980) allowing discussion of past changes in protist communities and in BB landscapes through time. These data are finally discussed in the light of reanalysis of regional precipitation signals (by modelling approaches based on NOAA data), instrumental data (nutrient concentrations) as well as historical chronicles (land-use practices and industrial-mining-war pressures) to tackle the main forcing factors responsible for coastal ecosystem transformations. We especially highlight the strong pressure of agriculture practices on trophic changes and degradation of BB’s water quality, as observed through coastal observatory series (IUEM, REPHY) with the recrudescence of toxic algal blooms since the 1980’s.
How to cite: Valero, C., Penaud, A., Vidal, M., Schmidt, S., Dessandier, P.-A., Goubert, E., Glemarec, E., Brigode, P., Bosseboeuf, L., Paulet, Y.-M., Liorzou, C., Revillon, S., Niass, N. C., Ailliot, P., Derrien, J.-M., Lambert, C., and Siano, R.: Past trajectory of a socio-ecosystem at the land-sea interface: the case of the northern watersheds of the Bay of Brest over the last 150 years, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-17787, https://doi.org/10.5194/egusphere-egu25-17787, 2025.
Hydrological connectivity is crucial for the mobilization, transport, and transformation of nitrate, but quantifying it at the catchment scale remains challenging, especially when capturing the spatial features that influence hydrological transport. We address this challenge by leveraging SMAP-Hydroblocks (Vergopolan et al., 2021), a high-resolution soil moisture dataset, to explore spatial soil moisture patterns as proxies for hydrological connectivity by predicting stream nitrate concentrations. We simulated daily nitrate concentrations across nine U.S. catchments with diverse land cover and concentration-discharge (C-Q) relationships using a multi-branch deep learning. We trained the model on discharge time series as an aggregated measure of hydrological connectivity, soil moisture spatial patterns to account spatial heterogeneities that influence hydrological connectivity, height above the nearest network maps as spatial flopath indicator and static proxies of nitrogen sources (nitrogen surplus and fraction of urban areas of catchments).
Our model achieved robust performance, with a median Nash-Sutcliffe Efficiency (NSE) of 0.63 and a median Kling-Gupta Efficiency (KGE) of 0.74 across the test period, outperforming traditional C-Q relationship models. Explainable AI (XAI) techniques revealed that spatial patterns of soil moisture contribute significantly to nitrate predictions, accounting for 30% of feature importance on average. Excluding these patterns decreased model accuracy by 14%. Explainable AI (XAI) methods revealed distinct hydrological responses across catchments: in catchments with positive C-Q patterns, spatial soil moisture patterns amplified nitrate transport during wet periods, while discharge dilution effects are more important in catchments with negative C-Q relationships. Attention maps highlighted near-stream zones as critical areas for predicting nitrate transport, reflecting their dominant role in hydrological connectivity and nitrate dynamics.
This study demonstrates the potential of integrating deep learning, XAI, and remote sensing products to quantify hydrological connectivity and nitrate dynamics. These findings provide new insights into the spatial and temporal variability of nitrate transport across catchments and a framework for improving water quality management.
Vergopolan, N., Chaney, N. W., Pan, M., Sheffield, J., Beck, H. E., Ferguson, C. R., Torres-Rojas, L., Sadri, S., & Wood, E. F. (2021). SMAP-HydroBlocks, a 30-m satellite-based soil moisture dataset for the conterminous US. Scientific Data, 8(1), 1. https://doi.org/10.1038/s41597-021-01050-2
How to cite: Saavedra, F., Vergopolan, N., Musolff, A., Merz, R., Winter, C., Wang, Z., and Tarasova, L.: From Soil Moisture Patterns to Hydrological Connectivity: An Explainable AI Approach for Nitrate Modeling, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-20591, https://doi.org/10.5194/egusphere-egu25-20591, 2025.
Effective water quality management at the catchment scale requires robust tools to predict pollutant loads under diverse environmental conditions. This study introduces an innovative methodology employing machine learning models to estimate phosphorus (P) concentrations and loads in small catchments contributing to the Northeast Arm of Lake Champlain (NALC). The region is characterized by agricultural land use and dynamic hydrological conditions. Current challenges include the lack of monitoring data for small direct drainage streams and the high uncertainty in existing P load estimates.
To address these gaps, we propose a dual modeling framework using Random Forest (RF) and Long Short-Term Memory (LSTM) models. Both models will be trained and validated using project-specific monitoring data, alongside extensive datasets from the USGS and regional monitoring programs. RF models, known for their interpretability and efficiency, will quantify predictor variable importance and generate insights into the key drivers of P loading. Complementarily, LSTM models, capable of capturing complex temporal dynamics, will provide high-resolution predictions of daily P loads and concentrations.
Our methodology highlights innovative monitoring strategies, including the deployment of stream gauging and water sampling stations at representative sites, capturing flow rates and concentrations of total phosphorus (TP), total dissolved phosphorus (TDP), and total suspended solids (TSS). These observations will be integrated into the machine learning framework, allowing a targeted validation of the model results. Preliminary analyses indicate disproportionately high P loading in small agricultural watersheds, underscoring the need for targeted interventions informed by reliable model predictions.
Expected outcomes of this study include the identification of source areas and processes driving P and sediment transport, as well as validated machine learning tools capable of estimating loads in ungauged basins. These models are designed to accommodate future scenarios of land use and climate change, providing resource managers with actionable insights to design effective mitigation strategies. By integrating high-resolution empirical data with state-of-the-art machine learning techniques, this work advances the understanding and management of nutrient dynamics at the catchment scale.
How to cite: Kiesel, J., Braun, D., and Vaughan, M.: Leveraging Machine Learning to Enhance Water Quality Predictions in Small Agricultural Streams , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-19410, https://doi.org/10.5194/egusphere-egu25-19410, 2025.
