Knowledge discovery from high-frequency stream nitrate concentrations: hydrology and biology contributions

Statistical evidence of recharge and supply controlling nitrate variability at springs discharging from the upper Floridan Aquifer

Over the last several decades, rising nitrate concentrations in springs discharging from north Florida's karstic Upper Floridan Aquifer have coincided with proliferation of algae in Florida spring runs and subsequent ecosystem degradation. As agriculture and development are primary contributors to groundwater nitrate and are predicted to continue expanding, understanding unique contributions and transmission pathways of nitrate pollution is vital to restoring impaired spring ecosystems. In this study, we use statistics and signal processing to analyze continuous nitrate timeseries data collected over five years at four north Florida springs. We quantified a significant, low-frequency annual signal in nitrate concentrations superimposed on increasing nitrate trends. We show nitrate concentrations at springs increase during the rainy season, potentially in response to recharge and seasonal fertilizer application. Thus, we suggest seasonal fluctuations observed in nitrate concentrations are caused by increased recharge of nutrient-rich soil waters through fractures that deliver water on relatively short timescales to conduits during the rainy season. We further speculate the steady, monotonically increasing concentration is maintained by accumulation of Nitrogen as slow flow through matrix porosity through the remainder of the year. Seasonal nitrate concentrations resulting from flow through karst aquifers may thus be poorly simulated using equivalent porous media models that are increasingly being used for nutrient management, because they do not capture heterogenous flow and transport dynamics.

Distance-based clustering challenges for unbiased benchmarking studies

Benchmark datasets with predefined cluster structures and high-dimensional biomedical datasets outline the challenges of cluster analysis: clustering algorithms are limited in their clustering ability in the presence of clusters defining distance-based structures resulting in a biased clustering solution. Data sets might not have cluster structures. Clustering yields arbitrary labels and often depends on the trial, leading to varying results. Moreover, recent research indicated that all partition comparison measures can yield the same results for different clustering solutions. Consequently, algorithm selection and parameter optimization by unsupervised quality measures (QM) are always biased and misleading. Only if the predefined structures happen to meet the particular clustering criterion and QM, can the clusters be recovered. Results are presented based on 41 open-source algorithms which are particularly useful in biomedical scenarios. Furthermore, comparative analysis with mirrored density plots provides a significantly more detailed benchmark than that with the typically used box plots or violin plots.

Explainable AI Framework for Multivariate Hydrochemical Time Series

The understanding of water quality and its underlying processes is important for the protection of aquatic environments. With the rare opportunity of access to a domain expert, an explainable AI (XAI) framework is proposed that is applicable to multivariate time series. The XAI provides explanations that are interpretable by domain experts. In three steps, it combines a data-driven choice of a distance measure with supervised decision trees guided by projection-based clustering. The multivariate time series consists of water quality measurements, including nitrate, electrical conductivity, and twelve other environmental parameters. The relationships between water quality and the environmental parameters are investigated by identifying similar days within a cluster and dissimilar days between clusters. The framework, called DDS-XAI, does not depend on prior knowledge about data structure, and its explanations are tendentially contrastive. The relationships in the data can be visualized by a topographic map representing high-dimensional structures. Two state of the art XAIs called eUD3.5 and iterative mistake minimization (IMM) were unable to provide meaningful and relevant explanations from the three multivariate time series data. The DDS-XAI framework can be swiftly applied to new data. Open-source code in R for all steps of the XAI framework is provided and the steps are structured application-oriented.

Temporal hydrochemical dynamics of the River Wensum, UK: Observations from long-term high-resolution monitoring (2011-2018)

In 2010, the UK government established the Demonstration Test Catchment (DTC) initiative to evaluate the extent to which on-farm mitigation measures can cost-effectively reduce the impacts of agricultural water pollution on river ecology whilst maintaining food production capacity. A central component of the DTC platform was the establishment of a comprehensive network of automated, web-based sensor technologies to generate high-temporal resolution (30 min) empirical datasets of surface water, groundwater and meteorological parameters over a long period (2011-2018). Utilising 8.9 million water quality measurements generated for the River Wensum, this paper demonstrates how long-term, high-resolution monitoring of hydrochemistry can improve our understanding of the complex temporal dynamics of riverine processes from 30 min to annual timescales. This paper explores the impact of groundwater-surface water interactions on instream pollutant concentrations (principally nitrogen, phosphorus and turbidity) and reveals how varying hydrochemical associations under contrasting flow regimes can elicit important information on the dominant pollution pathways. Furthermore, this paper examines the relationships between agricultural pollutants and precipitation events of varying magnitude, whilst demonstrating how high-resolution data can be utilised to develop conceptual models of hydrochemical processes for contrasting winter and summer seasons. Finally, this paper considers how high-resolution hydrochemical data can be used to increase land manager awareness of environmentally damaging farming operations and encourage the adoption of more water sensitive land management practices.

Hysteresis analysis of nitrate dynamics in the Neuse River, NC

Anthropogenic activities have caused N saturation in many terrestrial ecosystems. The transfer of nutrients and sediments to freshwater environments has resulted in water quality impairments including eutrophication, increased turbidity, ecosystem acidification, and loss of biodiversity. Storm events account for the transport of a large proportion of nutrients and sediments found in watersheds on an annual basis. To implement effective water-quality management strategies, the importance of surface and subsurface flow paths during storm events and low flow conditions need to be quantified. The increased availability of optical in-situ sensors makes high-frequency monitoring of catchment fluxes practical. In this study, we present a high-resolution nitrate monitoring record over a 10-year period in the Neuse River Basin near Clayton, North Carolina. The relationship between discharge and nitrate concentration for 365 storm events are categorized into hysteresis classes that indicate different transport mechanisms into the river. Storm events over the entire period of this study are divided between clockwise, counter-clockwise, and complex hysteresis patterns, indicating multiple nitrate flow paths during different seasons and years. Logistic regression of a suite of environmental variables demonstrates that antecedent soil moisture is a significant factor in determining the storm hysteresis class, with the odds of counter-clockwise hysteresis increasing by 10.3% for every 1 percentage point increase in the soil moisture. There is also an overlying seasonal effect, which indicates that dry soil conditions and frequent small storms during summer leads to greater nitrate transport on the rising limb, in contrast to slower, groundwater-driven inputs during the rest of the year.

Spatially distributed hydro-chemical data with temporally high-resolution is needed to adequately assess the hydrological functioning of headwater catchments

We demonstrated the great value of spatially distributed and temporally high-resolution hydro-chemical data to enhance knowledge about the intra-catchment variability of flow processes and the runoff composition of individual storms in a tropical alpine (Páramo) ecosystem. In this study, water sources (rainfall, spring water, and water from soil layers of Histosols and Andosols) and nested streams were sampled bi-weekly (2013-2014), including three storm high-resolution events (5-240 min). Water samples were analyzed for 14 tracers including electrical conductivity (EC) and rare earth trace elements and used as input to perform End-Member Mixing Analysis (EMMA). End-members identified for the outlet could explain the hydrological behavior of four out of the five tributaries, indicating similar hydro-geochemical processes and geomorphic features within the catchments. The runoff source contributions of the individual sub-catchments varied among (e.g. Andosols ~40% in tributaries and ~25% at the outlet) and within storm events (e.g. Histosols 15% higher in small peak discharge event), indicating a time-variable composition of streamflows. The latter was also reflected by the interaction of different sources and the chronology of flow paths in EMMA-space, evidencing a faster connectivity with hillslopes in the upper sub-catchments compared to the lower sub-catchments. We found counter-clockwise hysteresis patterns of storms in the lower catchments and clockwise hysteresis loops in the upper catchments. The latter bi-directionality can be related to lower slopes, wider riparian areas and the higher proportion of Histosols in the lower catchments compared to the upper sites.