Displays

In the age of big data, hydrologic studies contain more sites, longer and more resolute simulated and observed timeseries, and finer resolution spatial data than ever before.  This growth in capabilities to collect and generate data represents a tremendous opportunity for hydrologic science, but can challenge creating and presenting figures that summarize this information in succinct, interpretable, and meaningful ways.  To address this challenge, this presentation reviews several plotting approaches focused on synthesis of large hydrologic and environmental datasets from across the literature.  We highlight plots that can be used to visualize multi-dimensional spatial and temporal modeling and observational data, to synthesize patterns, to highlight outliers, and above all to convey key messages.  Building on these different types of plots, we highlight a set of best practices for how we as a community can create effective visualizations that synthesize large datasets of a variety of types in scientific presentations and publications.

How to cite: Kelleher, C. and Braswell, A.: Plotting for Synthesis of Large Datasets, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-1595, https://doi.org/10.5194/egusphere-egu2020-1595, 2020.

Rainfall-runoff models are widely used tools in catchment hydrology. Their evaluation is mostly based on comparing observed and simulated discharge values and various statistical objective functions have been proposed to evaluate the agreement between these time series. However, model evaluations that are based on statistical objective functions often does not provide the modeller with much insight on why the model fails to represent the hydrology of the real-world system.  Other, hydrologically meaningful indices or signatures have been proposed instead that quantify the hydrologic response characteristics of the catchment. They can also be regionalized and thus provide a potential opportunity for model evaluation in ungauged basins.

Our study investigates how to best integrate hydrological signatures in an objective function for model evaluation to shift the focus of objective functions to evaluate basic hydrological functions of catchments. We propose a signature-based hydrologic efficiency metric that can be derived from locally observed or regionalised hydrologic signatures.  The metric improves upon the Kling-Gupta Efficiency (KGE) metric by replacing its three components with hydrologic signatures characterising the water balance (or bias), the damping (or variance) and the timing of flows (or correlation). Additionally, we use these hydrologic signatures with the physical characteristics (i.e. catchment attributes) in some regionalization approaches such as linear, nonlinear regression and random forests for streamflow predictions in ungauged catchments. We test our ideas on a large and diverse sample of 582 UK catchments using the CAMELS-GB dataset and show that the performance of the proposed metric works well.

How to cite: Kiraz, M., Wagener, T., and Coxon, G.: Integration of Hydrologic Signatures for Model Evaluation in Gauged and Ungauged Catchments, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-7487, https://doi.org/10.5194/egusphere-egu2020-7487, 2020.

Hydrological models are conventionally evaluated in terms of their response surface or likelihood surface constructed with the model parameter space. To evaluate models as hypotheses, we developed the method of Flux Mapping to construct a hypothesis space based on model process representation. Here we defined the hypothesis space based on dominant runoff generating mechanisms, and acceptable model runs are defined as total simulated flow with similar (and minimal) model error simulated by distinct combinations of runoff components. We demonstrate that the hypothesis space in each modeling case is the result of interplay between the factors of model structure, parameter sampling, choice of error metric, and data information content. The aim of this study is to disentangle the role of each factor in this interplay. We used two model structures (SACRAMENTO and SIMHYD), two parameter sampling approaches (small samples based on guided-search and large samples based on Latin Hypercube Sampling), three widely used error metrics (NSE, KGE, and WIA — Willmott’s Index of Agreement), and hydrological data from a range of Australian catchments. First, we characterized how the three error metrics behave under different error regimes independent of any modeling. We then conducted a series of controlled experiments, i.e. a type of one-factor-at-a-time sensitivity analysis, to unpack the role of each factor in runoff simulation. We show that KGE is a more reliable error metric compared to NSE and WIA for model evaluation. We also argue that robust error metrics and sufficient parameter sampling are necessary conditions for evaluating models as hypotheses under uncertainty. We particularly argue that sampling sufficiency, regardless of the sampling strategy, should be further evaluated based on its interaction with other modeling factors determining the model response. We conclude that the interplay of these modeling factors is complex and unique to each modeling case, and hence generalizing model-based inferences should be done with caution particularly in characterizing hydrological processes in large-sample hydrology.

How to cite: Khatami, S., Peel, M., Peterson, T., and Western, A.: Characterizing dominant hydrological processes under uncertainty: evaluating the interplay between model structure, parameter sampling, error metrics, and data information content, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-11716, https://doi.org/10.5194/egusphere-egu2020-11716, 2020.

Conceptual Rainfall-Runoff (CRR) models are widely used for runoff simulation, and for prediction under a changing climate. The models are often calibrated with only a portion of all available data at a location, and then evaluated independently with another part of the data for reliability assessment. Previous studies report a persistent decrease in CRR model performance when applying the calibrated model to the evaluation data. However, there remains a lack of comprehensive understanding about the nature of this ‘low transferability’ problem and why it occurs. In this study we employ a large sample approach to investigate the robustness of CRR models across calibration/validation data splits. Specially, we investigate: 1) how robust is CRR model performance across calibration/evaluation data splits, at catchments with a wide range of hydro-climatic conditions; and 2) is the robustness of model performance somehow related to the hydro-geo-climatic characteristics of a catchment? We apply three widely used CRR models, GR4J, AWBM and IHACRE_CMD, to 163 Australian catchments having long-term historical data. Each model was calibrated and evaluated at each catchment, using a large number of data splits, resulting in a total of 929,160 calibrated models. Results show that: 1) model performance generally exhibits poor robustness across calibration/evaluation data splits; 2) lower model robustness is correlated with specific catchment characteristics, such as a higher runoff skewness, lower aridity and runoff coefficient. These results provide a valuable benchmark for future model robustness assessments, and useful guidance for model calibration and evaluation.

How to cite: Zheng, F., Guo, D., Gupta, H., and Maier, H.: On the Robustness of Conceptual Rainfall-Runoff Models to Calibration and Evaluation Dataset Splits Selection: A Large Sample Investigation, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-3205, https://doi.org/10.5194/egusphere-egu2020-3205, 2020.

The study aims to propose and test the performance of a group of techniques for transposing rainfall-runoff model parameters to ungauged catchments, especially adapted to the semi-distributed structure (the catchment is split into different altitude zones) of the HBV-based TUWien model.

The methods are tested for two large, but deeply different, datasets: the first is a very densely gauged set of more than 200 catchments across Austria, while the second refers to more than 500 US watersheds (part of the CAMELS dataset) covering most of the country, including wider variety of hydrological conditions and catchment characteristics.

The potential of the semi-distributed structure is fully exploited: first in the model calibration, where, differently from the typical application of the model, the parameters controlling the runoff generation are allowed to vary over the different elevation zones.

Secondly, in the regionalisation procedure, the parameters of each specific altitude zone in any ungauged catchment are estimated based on the parameters obtained for the same altitude zones of the donors. The rationale is to implement a procedure that operates at sub-basin level, in order to have a better simulation of the different hydrological processes taking place at different altitudes.

The set of regionalisation approaches includes both i) “parameters averaging”, where each parameter is obtained as a weighted (according to donors’ similarity) average of the parameters of the donor catchments (independently from each other) and ii) “output averaging”, where the model is run over the ungauged basin using the entire set of parameters of each donor basin and the simulated outputs are then averaged to estimate the target simulated streamflow.

The measure of similarity needed for implementing the regionalisation procedure is of course applied at sub-basin scale, testing geo-morphological and climatic catchment descriptors characterising the elevation bands. One of the main focus is the study of such similarity in order to asses which attributes are more influential at different altitudes.

The performance of the proposed approaches and similarity measures is assessed by jack-knife cross-validation against the observed daily runoff for all the study catchments.

Finally, the resulted regionalisation efficiencies are compared to those obtained by applying the same methods with the typical lumped calibration-regionalisation procedure, thus assessing the potential of the semi-distributed regionalised parameterisation.

How to cite: Neri, M., Toth, E., and Parajka, J.: Exploring elevation zone similarity in large case studies for the semi-distributed regionalisation of the HBV model parameters, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-658, https://doi.org/10.5194/egusphere-egu2020-658, 2020.

The parameterization of hydrological models in ungauged catchments remains challenging. However, the increased availability of large-sample data sets in recent years provides new opportunities for regionalization. In this study, we use the CAMELS dataset and the HBV model to simulate daily runoff in nearly 600 catchment areas in the United States. In a first step,  a lower and an upper benchmark were derived to obtain an approximation for how poor or how good runoff simulation could get in each of the catchments. For the upper benchmark the HBV model was calibrated, and the calibrated parameter values ​​were related to catchment characteristics. To account for parameter uncertainty, 100 independent calibrations were performed, and then median efficiency values ​​ were used for further analyses. For the lower benchmark the HBV model was run for 1000 randomly selected parameter sets, and median efficiency values were again used for further analyses. In a second step, each catchment was treated as ungaued and its parameter values were estimated by multiple regionalization methods. For each regionalization method donor catchments were selected based on a certain criterion including spatial proximity, similarity of hydrological signatures or attribute similarity. Additionally, we tested the added value of single discharge observations, which could be collected during short field visits. Furthermore, to analyze the theoretical limits of regionalization in general, the best three available donors of each receiver catchment were directly used to run simulations. All regionalization approaches were evaluated based on their relative performance with regard to the upper and lower benchmark.
First results indicated that the use of an ensemble of parameter sets calibrated in one of the gauged catchments leads to clearly better simulations than the use of randomly selected parameter values. Using the best three donor catchments resulted in nearly as good simulations as the upper benchmark, showing that regionalization has a high potential as long as we find a way to select these most suitable donors. The regionalization approach coming closest to the upper benchmark was based on a combination of spatial proximity and the use of single discharge measurements. Yet, there was still a considerable gap to the performance of using the best three donors. Despite the potential of regionalization demonstrated in this study, there still remains the challenge to find more reliable ways to link the hydrological functioning of catchments with the similarity of model parameter values.

How to cite: Vis, M., Pool, S., and Seibert, J.: Parameter values for ungauged catchments - comparing regionalization approaches using large-sample hydrology, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-17393, https://doi.org/10.5194/egusphere-egu2020-17393, 2020.

Design flood estimation in data-poor regions is a fundamental task in hydrology. In this paper, we propose a regional flood frequency analysis approach to estimate design floods anywhere on the global river network. This approach involves two stages: (i) clustering global gauging stations into subareas by a K-means model based on twelve globally available catchment descriptors and (ii) developing a regression model in each subarea for design flood estimation using the same descriptors. Nearly 12,000 discharge stations globally were selected for model development and a benchmark global index-flood method was adopted for comparison. The results showed that: (1) the proposed approach achieved the highest accuracy for design flood estimation when using all catchment descriptors for clustering; and the regression model accuracy improved by considering more descriptors in model development; (2) a support vector machine regression showed the highest accuracy among all regression models tested, with relative root mean squared error of 0.67 for mean flood and 0.83 for 100-year return period flood estimations; (3) 100-year return period flood magnitude in tropical, arid, temperate, continental and polar climate zones could be reliably estimated with relative mean biases of -0.18, -0.23, -0.18, 0.17 and -0.11 respectively by adopting a 5-fold cross-validation procedure; (4) the proposed approach outperformed the benchmark index-flood method for 10, 50 and 100 year return period estimates; We conclude that the proposed RFFA is a valid approach to generate design floods globally, improving our understanding of the flood hazard, especially in ungauged areas.

How to cite: Zhao, G., Bates, P., Neal, J., and Pang, B.: An improved regional flood frequency analysis approach at the global scale, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-11288, https://doi.org/10.5194/egusphere-egu2020-11288, 2020.

Characteristics and process controls of statistical moments of annual maximum peak discharges, including the mean annual flood (MAF), the coefficient of variation (CV) and the coefficient of skewness (CS), are analyzed for flood series in Europe. The data set consists of observations from 2370 catchments with an average record length of 48 years. The controls are identified by investigating dependencies between the flood moments and catchment area, flood seasonality, climate and catchment characteristics in five regions. The covariates providing the most explanatory power for within-region variability are identified in a regression framework. Preliminary results indicate: MAF and CV are strongly correlated with hydroclimatic catchments characteristics, and to a lesser degree with topography and land use. In the Atlantic region, precipitation is the most important control on the spatial patterns of MAF and CV; in the Mediterranean it is precipitation and aridity; and in Northeastern Europe it is air temperature.

How to cite: Lun, D., Viglione, A., Komma, J., Bertola, M., Parajka, J., and Blöschl, G.: Characteristics and process controls of statistical flood moments in Europe - a data based analysis, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-16969, https://doi.org/10.5194/egusphere-egu2020-16969, 2020.

Hydrology and especially hydrological models often treat catchments as if they were leaky buckets. But, do we find such catchments in the real world or is this just a convenient simplification? Moreover, if we find them, what attributes allow these catchments to show such a simple behavior? To study this, we look at time series of 27 years for 90 catchments in Hesse, Germany, which includes droughts and years of abundant precipitation. In addition, the state Hesse provides a wide range of catchment attributes like geology, soils and land use, while still having a relatively similar climate. Using discharge, evapotranspiration and precipitation, we calculate the cumulative storage change for all years separately and use it as a proxy for the storage. We group the 90 catchments by the complexity of their storage-discharge relationship, which we define as how good the relationship can be modelled by an exponential function. We find that climate and physical attributes of the catchments seem to have similar influence on the overall complexity of the storage-discharge relationship. However, we could also identify catchments that depict consistent behavior, mostly independent of climate. Those catchments either behave always complex or always simple in all the years considered. They differ in their permeability, conductivity, geology, soil and to a lesser extent their shape. We show that bucket like catchments exist in the real world and that they can be found by looking for oval catchments with good permeability in regions of igneous geology and clay silt soil texture.

How to cite: Jehn, F. U., Breuer, L., Kraft, P., Bestian, K., and Houska, T.: Simple catchments and where to find them: The storage discharge relationship as a proxy for catchment complexity, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-4540, https://doi.org/10.5194/egusphere-egu2020-4540, 2020.

Over the last decade or so many studies of the hydrologic characteristics of basins have been driven by the desire to develop models that enable prediction of particular signatures, such as baseflow and Base Flow Index (BFI), in ungauged basins (PUB). These studies typically focus on understanding how readily available mapped or remotely sensed data can be used to infer hydrologic signals. However, in the specific case of baseflow, there is a recognition that we still have a poor understanding of the relative influences of underlying hydrological processes at appropriate scales, particularly in anthropogenically impacted catchments. New opportunities are being offered to better understand relationships between BFI and various controls on baseflow through the production of large sample catchment datasets. Here we present the results of an analysis of one such large-sample dataset, CAMELS-GB, investigating the relative importance of different hydrogeological controls on baseflow, including factors such as: climatology; hydrogeology; geophysical catchment characteristics, e.g. soil characteristics and land cover; and, anthropogenic influences, e.g. discharge from reservoirs and from sewage treatment works (STWs), abstraction, and mains leakage.

CAMELS-GB consists of daily hydrometerorological time series for the period 1970-2015 and landscape, catchment and hydrogeological attributes for 671 catchments in Great Britain. Machine learning approaches, including random forest algorithms, are used to investigate the influence of catchment characteristics on BFI and to inform the selection of hydrologically reasonable parameters to quantify relationships using linear regression models. We describe how the regression models can be used to investigate and characterise the sensitivity of estimates of BFI to: i.) the underlying hydrogeological mapping; ii.) the spatial support scale of the analysis; and iii.) anthropogenic influences.   

How to cite: Bloomfield, J., Addor, N., Coxon, G., Gong, M., and Marchant, B.: Using the CAMELS-GB large-sample dataset to investigate controls on baseflow (BFI), EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-5509, https://doi.org/10.5194/egusphere-egu2020-5509, 2020.

The streamflow seasonality in mountain catchments is largely influenced by snow. However, a shift from snowfall to rain is expected in the future. Consequently, a decrease in snow storage and earlier snowmelt is predicted, which will cause changes in spring and summer runoff. The objectives of this study were to quantify 1) how inter-annual variations in snow storages affect spring and summer runoff, including summer low flows and 2) the importance of snowmelt in generating runoff compared to rainfall. The snow storage, groundwater recharge and streamflow were simulated for 59 mountain catchments in Czechia in the period 1980–2014 using a bucket-type catchment model. The model performance was evaluated against observed daily runoff and snow water equivalent. Hypothetical simulations were performed, which allowed us to analyse the effect of inter-annual variations in snow storage on seasonal runoff separately from other components of the water balance. This was done in the HBV snow routine using the threshold temperature T_T that differentiates between snow and rain and sets the air temperature of snowmelt onset. By changing the T_T, we can control the amount of accumulated snow and snowmelt timing, while other variables remain unaffected.

The results showed that 17-42% (26% on average) of the total runoff in study catchments originates as snowmelt, despite the fact that only 12-37% (20% on average) of the precipitation falls as snow. This means that snow is more effective in generating catchment runoff compared to liquid precipitation. This was documented by modelling experiments which showed that total annual runoff and groundwater recharge decreases in the case of a precipitation shift from snow to rain. In general, snow-poor years are clearly characterized by a lower snowmelt runoff contribution compared to snow-rich years in the analysed period. Additionally, snowmelt started earlier in these snow-poor years and caused lower groundwater recharge. This also affected summer baseflow. For most of the catchments, the lowest summer baseflow was reached in years with both relatively low summer precipitation and snow storage. This showed that summer low flows (directly related to baseflow) in our study catchments are not only a function of low precipitation and high evapotranspiration, but they are significantly affected by previous winter snowpack. This effect might intensify the summer low flows in the future when generally less snow is expected.

Modelling experiments also opened further questions related to model structure and parameterization, specifically how individual model procedures and parameters represent the real natural processes. To understand potential model artefacts might be important when using HBV or similar bucket-type models for impact studies, such as modelling the impact of climate change on catchment runoff.

How to cite: Jenicek, M. and Ledvinka, O.: Snowmelt contribution to seasonal runoff: Lessons learned from using a bucket-type model on a large set of catchments, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-5557, https://doi.org/10.5194/egusphere-egu2020-5557, 2020.

Abstract: Land use/land cover change (LULC) in glacial affected areas are driven by climate change and human activities. Monitoring and simulation of the spatial and temporal land cover changes in this special region provide scientific basis in understanding the natural environment, helping to reveal the impact of climate change and human activities on LULC. In this study, the Tianshan Mountains (TSM), located in the hinterland of Eurasia, were selected as the study area to investigate the LULC of the glacial affected areas. The relationship between LULC, human intervention and climate change on a large spatial scale were also analyzed. The LULC of the TSM in China for the past 35 years were analyzed using a dynamical change model, a landscape pattern index, a centroid transfer model, and geoinformation TUPU based on the land use data of 1980, 1990, 2000, and 2015. Results show that the areas of cultivated and built-up lands immensely increased by 45.87% and 187%, respectively. Correspondingly, the areas of bare land and ice and snow cover decreased by 27% and 38%, respectively. The land use change in the TSM was characterized by different stages, and high conversion rate and intensity were obtained from 2000 to 2015. The landscape change was mainly reflected in terms of the significant increase in the number of patches and the simplification and regularization of patch shapes. The spatial connectivity of different land use types increased. The temperature and precipitation in the region show an increasing trend, and the melting rate of ice and snow cover significantly accelerated. This study can help to achieve a dynamic LULC model to investigate the interacting influences of climate change and human activities in glacial affected areas.

How to cite: Wei, H., Xiong, L., Tang, G., Strobl, J., and Xue, K.: Impact of human intervention and climate change on the land cover change in the glacial affected area of the Tianshan Mountains, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-2138, https://doi.org/10.5194/egusphere-egu2020-2138, 2020.

Improving our understanding of hydrological processes beyond single catchments is important. Here we test wflow_sbm (simple bucket model) for modelling hydrology across different geographical areas (i.e. Europe, Africa).  Wflow_sbm is a conceptual bucket-style hydrological model based on simplified physical relationships. It uses kinematic wave surface and subsurface routing for lateral transport. The model setup and parameter estimation are fully automated based on global and regional data sources (like MERIT DEM, SoilsGrids, monthly MODIS LAI, global/regional land use) and includes anthropogenic influences like lakes and reservoirs and its management from HydroLAKES and GRanD databases. It makes use of scaling operators as applied in Multiscale Parameter Regionalization (MPR) to go from high resolution data sources to a ~1km2 or coarser resolution model.  

The model is tested at 1 km2 spatial and daily temporal resolution for different basins: Umealven (Sweden), Glomma (Norway), Mono river (Togo), the white Nile, Save River (Mozambique) and Incomati river using EOBS or CHIRPS rainfall and ERA5 derived temperature and potential evaporation forcing.  From these applications, it becomes apparent that the model can explain the measured discharge most of times reasonably well (KGE~0.4 and higher). The main factors controlling the performance are (quality of the) forcing, lateral hydraulic conductivity, rooting depth and reservoir/lake management.

How to cite: Weerts, A., van Verseveld, W., Eilander, D., Boisgontier, H., Haag, A., Hazenberg, P., and Imhoff, R.: Testing the distributed hydrological wflow_sbm concept across different geographical domains, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-10886, https://doi.org/10.5194/egusphere-egu2020-10886, 2020.

Advances in open data science serve large-scale model developments and, subsequently, hydroclimate services. Local river flow observations are key in hydrology but data sharing remains limited due to unclear quality, or to political, economic or infrastructure reasons. This presentation provides methods for quality checking openly accessible river-flow time series. Availability, outliers, homogeneity and trends were assessed in 21 586 time series from 13 data providers worldwide. We found a decrease in data availability since the 1980s, scarce open information in southern Asia, the Middle East and North and Central Africa, and significant river-flow trends in Africa, Australia, southwest Europe and Southeast Asia. We distinguish numerical outliers from high-flow peaks, and integrate all investigated quality characteristics in a composite indicator.

Some 5338 gauges from these river flow time series (> 10 years) were used in the evaluation of the Worldwide-HYPE (WWH) hydrological model at the global scale (half for calibration and half for independent validation), resulting in a median monthly KGE of 0.4. However, WWH performance varies widely spatially and with the target flow signature. The model performs best (KGE > 0.6) in Eastern USA, Europe, South-East Asia, and Japan, as well as in parts of Russia, Canada, and South America. It also shows overall good potential to capture flow signatures of monthly high flows, spatial variability of high flows, duration of low flows, and constancy of daily flow. Nevertheless, there remains large potential for model improvements and we suggest both redoing the parameter estimation and reconsidering parts of the model structure for the next WWH version.

References:

Crochemore, L., Isberg, K., Pimentel, R., Pineda, L., Hasan, A., Arheimer, B., (2019). Lessons learnt from checking the quality of openly accessible river flow data worldwide. Hydrological Sciences Journal. https://doi.org/10.1080/02626667.2019.1659509

Arheimer, B., Pimentel, R., Isberg, K., Crochemore, L., Andersson, J. C. M., Hasan, A., Pineda, L., (accepted). Global catchment modelling using World-Wide HYPE (WWH), open data and stepwise parameter estimation. Hydrology and Earth System Sciences Discussions. In press. https://doi.org/10.5194/hess-2019-111

How to cite: Arheimer, B., Crochemore, L., Pineda, R., Isberg, K., Pineda, L., Hasan, A., and Andersson, J.: Lessons learnt from quality-checking observed and simulated river flow data worldwide, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-8292, https://doi.org/10.5194/egusphere-egu2020-8292, 2020.

Aluminum is toxic to most aquatic and terrestrial organisms. Increased Al concentrations in soils and freshwaters are a direct result of human activity, via increases in acid deposition. Elevated Al concentrations pose a wide variety of threats to ecosystems and society, from causing human neurotoxicity, reducing carbon sequestration in forests, threatening biodiversity, and increasing the cost of water treatment. Freshwater aluminium concentrations increased across Europe and North America between the 1960s and 1990s, predominantly due to ecosystem acidification. Following acidic deposition reduction legislation enacted in the 1990s, the problems of acidification and increased freshwater aluminium concentrations were considered solved. However, recently and unexpectedly, Sterling et al. identified aluminum concentrations to be increasing across North America and Scandinavia. Sterling et al. proposed a conceptual model suggesting these widespread increases in freshwater aluminium concentrations resulted from a hysteresis of base cation and dissolved organic carbon (DOC) response to decreasing acidic deposition, where base cation increase is slow compared to that of DOC, resulting in elevated freshwater aluminium concentrations. This process can be exacerbated by further increases in DOC due to increasing global surface temperatures. The Sterling et al. conceptual model is supported by prior work by Weyhenmeyer et al. (2019, Scientific Reports) and Monteith et al., 2007, Nature) who identified widespread decreasing calcium and increasing DOC concentrations. In this study, we aim to validate the Sterling et al. model and identify if it is generalizable to other regions with decreasing calcium and increasing DOC trends, irrespective acidification status. Additionally, we aim to characterize other regions across the globe which are at risk of elevated aluminium concentrations. To fulfill these research goals, we compiled a large-sample water chemistry database from existing national and global datasets – GloFAD (Global Freshwater Acidification Database). The database is comprised of over 11 million unique samples spanning nearly 286,000 sites located between Antarctica and Russia, 18 years (2000 to 2019), and 40 water chemistry parameters. Preliminary analysis shows that aluminium is increasing in 27% to 71% of sites, dependent on the species, base cations are decreasing for 62% to 70% of sites, freshwater organic carbon is increasing for 58% to 64% of sites, and water temperature is increasing in 61% of sites. Increasing dissolved aluminium trends are strongly significantly correlated with decreasing base cation trends (calcium τ = -0.71 and magnesium τ = -0.59, α < 0.05) but not with DOC concentrations (τ = -0.08). The lack of correlation with DOC indicates that drivers of increasing aluminium trends may differ based on the acidification status of the watershed and that regional models of freshwater aluminium chemistry may not be globally applicable. The widespread decreasing base cation trends and strong correlation between decreasing base cation and increasing aluminium trends indicates that increasing aluminium concentrations may become more widespread, posing a threat to aquatic and terrestrial organisms, potentially including humans, reducing carbon sequestration in forests, threatening biodiversity, and increasing water treatment costs.

How to cite: Rotteveel, L. and Sterling, S.: Identifying global trends and drivers of freshwater aluminium concentrations using GloFAD (Global Freshwater Acidification Database), EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-10854, https://doi.org/10.5194/egusphere-egu2020-10854, 2020.