Assessing the spatio-temporal dynamics and environmental impacts in peri-urban stream systems

Stream water resources and ecosystems are at threat worldwide from the ever-increasing pressure of urban and agricultural expansion, as well as climate change. Furthermore, the ongoing green transition of the world economy has led to emerging concerns of unintentional negative consequences with respect to freshwater biodiversity. These different stressors and how they contribute to the degradation of stream quality and ecology is especially complex to com-prehend in peri-urban landscapes. These landscapes are indeed characterized by a sprawl of urban, industrial and agricultural activities to the detriment of more natural areas, and the inherent variability of these catchments results in highly dynamic hydromorphological and water quality stressors.

Getting a better understanding of these conditions, stressor dynamics and the effects of underlying driving processes on ecological health is therefore an im-portant priority. Dedicated measurements and in-situ monitoring are standard approaches to tackle this knowledge gap, but to this day are still practically and economically constrained in peri-urban settings due to their dynamic na-ture and high variability. Simulation models can be used to advance the under-standing of processes affecting stream water quantity and quality, and enable the design of more efficient monitoring programs. But while numerous hydro-logical and water quality models exist, specific applications to peri-urban streams, for decision-making and mitigation strategies are still limited.

The main objective of this PhD was therefore to further the understanding of spatio-temporal variations affecting water quantity and quality in peri-urban stream systems, and how these parameters can potentially impact ecological quality. In order to fulfil this scope, two main field investigations (reach and catchment scales) were carried out to evaluate the variation in discharges from contaminant sources and variability of physico-chemical stream conditions, with sampling on a monthly to bi-monthly basis over a year. An ecological quality assessment was carried using traditional (benthic macroinvertebrate) and novel (meioinvertebrate, specifically nematodes) bioindicators. Additionally, an integrated model (water quantity and quality, i.e. water temperature, dissolved oxygen (DO), and macronutrients) was implemented to obtain better insights into the dynamics of peri-urban stream systems. This model was developed using a System Dynamics approach for transparency purposes, to enable future stakeholder engagement practices and facilitate its potential transferability to other catchments at a later stage.

Both field studies revealed high seasonal variations of flow stemming from the numerous pathways that contribute to peri-urban stream discharges. Notably, the reach scale investigation (Mølle Stream) highlighted alterations of the groundwater baseflow and how different urban features and former anthropogenic modifications (retainment wall, drains) resulted in the creation of complex pathways that drive important spatio-temporal variations of contaminated groundwater and other discharges, and ultimately, variation in chemical concentrations in the receiving stream.

A tracer experiment combined with a ground truth dataset were performed in Mølle Stream in order to investigate the potential use of hyperspectral drone images in future tracer tests. The experiment also investigated variations in concentrations and mixing processes in streams. Of particular relevance for this study, the distance between a discharge entry point or a contaminant source and fully mixed conditions (i.e. the mixing length) were assessed. The comparison between the tracer experiment, the reach scale investigation and a dedicated steady-state model for the estimation of the mixing length were in reasonable agreement. The results showed that the variations of tracer concentrations (and in general any dissolved substances) are strongly dependent on sample location, up to the point of fully mixed conditions. The mixing length is affected by the flow variation and mixing process and hence should be considered when applying a water quality model or when sampling in relation to compliance.

The field investigation at the catchment scale (Usserød Stream) showed the strong variability of physico-chemical conditions in a peri-urban stream, and underlined the dynamic and variable contributions of different land-use areas. Wastewater treatment plant (WWTP) effluents were important contributors of nutrients (N, P), but significant contributions from agricultural lands were also captured during specific periods.

The model co-developed during this PhD was applied in the Usserød Stream catchment and successfully captured flow, depth and temperature variations. Notably, it showed the effect of enhanced heterotrophic respiration in summer which resulted in lower oxygen concentrations, possibly driven by (1) dissolved organic carbon emanating from different agricultural or urban areas, or (2) the enhanced settling and degradation of organic particles provoked by a combination of shallow depth and interception by water plants. A potential remobilization of phosphorus from the streambed was also highlighted, through uncertainty analysis results combined with the rich dataset collected on site. This process may become a significant diffuse source of nutrients, if others (e.g. wastewater treatment plant effluent) were eliminated.

The temporal variations of some physico-chemical conditions (specifically DO, temperature and Biological Oxygen Demand) were seen as an explanatory factor for the poor to moderate ecological status recorded at the sampling stations along the Usserød Stream (macroinvertebrate and fish bioindicators). Nevertheless, these variations were comingled with other potential stressors such as degraded physical habitat, lack of a source population for recolonization (macroinvertebrates) and chemical contamination. Indeed, the ecological assessment utilizing a nematode-based stress index (nemaSPEAR[%]), well suited to reveal chemical contamination, showed bad to moderate quality at some sampling stations where suspended particles (and thus solid-bound contaminant) concentrations were high, suggesting that chemical stressors are likely overlooked.

The aforementioned nemaSPEAR[%] index appears as a powerful bioindicator for environmental impacts and inherent dynamics in peri-urban stream systems. This indicator could discriminate different urban features such as wastewater treatment plant effluents and combined sewer overflows. Furthermore, it revealed potential benefits of these urban wastewaters which so far remained unnoticed in the development of green solutions. These solutions promote, for instance, more centralized and resource/energy-efficient wastewater treatment plants along with potential re-routing of effluent discharges, and could ultimately result in unforeseen negative impacts on ecological quality.
Overall, these findings highlight the need to continuously track the dynamic properties in peri-urban stream systems, ideally at different time scales (high frequency, short term, seasonal, long term) and taking advantage of the new-est available monitoring technologies. The collected data, in close combina-tion with simulation tools, will better support process understanding, the quantification of contributions from contaminant sources and different land-use and impacts in terms of both water and ecological quality.