Samenvatting

Abstract This study evaluated the effects of vegetation and internal water storage (IWS) on evapotranspiration (ET), hydrologic performance, and water quality in bioretention mesocosms. Twelve mesocosms were configured as planted/free-drainage (PF), planted/IWS (PI), unplanted/free-drainage (UF), and unplanted/IWS (UI). Daily ET was quantified by mass balance, and environmental drivers were analysed using linear mixed-effects models. ET was primarily controlled by atmospheric demand, with relative humidity (marginal R² = 0.35) and precipitation (0.29) explaining more variability than solar radiation or temperature (~0.26). Vegetation (the tested species was Salix integra) significantly increased ET, whereas IWS had no effect, indicating energy-limited conditions. However, IWS reduced overflow, especially in planted systems. Water quality results showed higher electrical conductivity in IWS-treated unplanted systems, while planted systems maintained lower values. Outflow pH remained stable. These findings highlight vegetation as the main driver of ET, while IWS enhances retention but may increase solute concentrations in the absence of vegetation. Keywords Bioretention; Evapotranspiration; Internal water storage; Nature-based solutions; Vegetation; Water quality INTRODUCTION Bioretention systems (BRs) are widely implemented as nature-based solutions for urban stormwater management, helping to reduce runoff, mitigate flooding, and support groundwater recharge (Boogaard et al., 2025). While these systems have traditionally been designed to maximize infiltration, evapotranspiration (ET) is an equally important process that contributes to the recovery of hydraulic capacity, enhances water retention, and mitigates urban heat island effects. ET is a major component of the BR’s water balance, returning 30–90% of captured rainfall to the atmosphere (Ouédraogo et al., 2022). It is commonly assumed that ET is primarily controlled by vegetation and soil moisture availability. Accordingly, internal water storage (IWS) has been proposed as a design strategy to sustain root-zone moisture and enhance ET between storm events. However, ET is also strongly influenced by atmospheric drivers, including solar radiation, temperature, air humidity, wind, and precipitation, which govern energy availability and evaporative demand. In shallow engineered substrates, rapid moisture depletion can limit transpiration (De Ville et al., 2024), yet it remains unclear whether ET in BRs is primarily constrained by water availability or by atmospheric conditions. Consequently, the relative importance of vegetation, IWS, and environmental drivers in controlling ET remains insufficiently quantified under realistic conditions. This study addresses this gap by evaluating how vegetation, IWS, and atmospheric drivers jointly influence ET dynamics in replicated mesocosm-scale BRs under outdoor conditions, providing new insights for climate-resilient system design. MATERIALS AND METHODS Twelve cylindrical 150 L mesocosms were used to evaluate the effects of vegetation (Plant species: Salix integra) and IWS on ET in BRs. Each mesocosm consisted of a layered profile (from bottom to top: 10 cm gravel, 30 cm sand, and 50 cm loamy soil mixture) and was equipped with a drainage outlet. In IWS treatments, the outlet height was elevated to create a perched water table. Four configurations were tested: planted/free drainage, planted/IWS, unplanted/free drainage, and unplanted/IWS, with three replicates each. Mesocosms were placed outdoors (June–September) and periodically weighed to estimate ET using a mass-balance approach, accounting for drainage. Soil moisture and temperature were monitored at three depths using sensors. Meteorological variables (air temperature, relative humidity, wind speed, and solar radiation) were recorded. Electrical conductivity and pH were measured from the effluents of the mesocosms. Linear mixed-effects (LME) models were applied to quantify the influence of environmental drivers and design factors on daily ET. RESULTS AND DISCUSSION Vegetation was the dominant control on ET, with planted mesocosms exhibiting substantially higher daily and cumulative ET than unplanted BRs (~5.2 vs ~1.7–2.1 mm d⁻¹). Mixed-effects models showed that ET variability was primarily driven by atmospheric conditions, with relative humidity (marginal R² = 0.35) and precipitation (0.29) explaining more variance than solar radiation or temperature (~0.26), indicating an energy-limited but demand-modulated system. Although IWS increased soil moisture at depth and cumulative ET amounts, this impact was not statistically significant in most LME models, confirming that water availability was not limiting. Soil moisture modelling results showed that IWS significantly influenced plant-available water content, while willows had no consistent effect on the content across depths. The present analysis indicates that, while IWS prevents moisture limitation, ET magnitude remains primarily controlled by atmospheric demand. Together, these results suggest that ET of BRs under these conditions is primarily energy-limited, with IWS ensuring moisture sufficiency but not increasing evaporative demand. However, IWS significantly improved hydrologic performance: overflow volumes decreased from 2.9 L in unplanted free-draining systems to 1.2 L in planted systems with IWS. This reduction reflects both increased ET in vegetated BRs and prolonged residence time under IWS conditions. Electrical conductivity (EC) of outflow was primarily influenced by IWS configuration, with BRs with IWS showing consistently higher EC than free-draining units, while plants had no significant independent effect. However, planted BRs with IWS showed lower cumulative solute outflow as expressed by EC multiplied by outflow volume. Overall, these findings show that vegetation enhances ET and mitigates water-quality impacts, while IWS primarily improves retention, underscoring the importance of integrating both features into BR design.