4.1 Evaluation of groundwater flow velocities and clogging
All the hydraulic experiments conducted at Kojetín experimental site, including several pumping tests and VSK infiltration, were analysed using MODGLOW-USG software (Panday et al., 2013). The transport of non-sorbing, non-reacting PPCPs from VSK well was simulated too.
The agreement between measurement and model result was achieved with the aquifer porosity of 15 %, which is the expected value when considering a fluvial aquifer. In accordance with the measurement, the model also predicted that the locality of the PZ5 well remained unaffected by transport from the VSK well.
Model calculated breakthrough time (3 – 5 days) is shorter than observed (8 – 12 days), even though the model longitudinal dispersivity was reduced to 1 m to minimise dispersion. This suggests that during the initial phase of the transport process, sucralose was degraded. The same process probably took place which caused the initial content of this substance to be zero.
Clogging of the VSK borehole was manifested by a huge difference in groundwater level response during the pumping and infiltration period. Throughout the proceeding pumping test, the water level in the VSK well decreased by 1 m, when 4 l/s were abstracted. The infiltration experiment was finished with the recharge rate of approximately 0.7 l/s and the water level increased by 4 m. The level difference between the VSK borehole and the casing observation probe increased too.
The persistently relatively small level difference between the infiltration well and the casing probe proves that the hydraulic resistance against infiltration arose only at the filter pack/aquifer interface. It is likely that physico-chemical processes caused by the seepage of oxygenated water into the originally anoxic environment, in combination with microbiological processes and sorption to precipitated Fe3+ compounds, contributed significantly to the clogging. It is evident from the obtained data that the artificial recharge of wastewater using the technology of infiltration wells will bring technical problems associated with a decrease in yield.
4.2 Evaluation of qualitative changes
Tab. 1 provides a complete overview of the occurrence of PPCP substances at the monitored objects in the period before and during the infiltration experiment. For substances that were analyzed in more detail, the mean and median content values are also given.
A comparison of both recipients of PPCPs (i.e. treated wastewater from the WWTP and water in the Morava River) reveals their similarity in terms of the range of detected components. Many other wastewater treatment plants are connected to Morava above the Kojetín location. A total of 80 PPCP substances with values above the detection limit appeared in at least one sample on the monitored objects.
The permanent occurrence exhibited 45 PPCP substances concerning treated wastewater and 29 substances concerning water in the Morava River. At the beginning of the VSK infiltration experiment, well PV5 already contained 17 PPCP substances, whilst wells CP1 and PZ1 contained 10 substances. The PV5 well is located in the proximity of the Morava River (Fig. 2). It follows that the river water negatively affects the adjacent fluvial aquifer and a non-negligible set of PPCP substances is “naturally” present in the groundwater unaffected by the infiltration experiment. During floods and due to the dam cascade the Morava River occasionally has effluent, or permanently in some of its parts. This causes the PPCP to transport from the river into fluvial sediments.
Table 1 Information on the presence of PPCP substances at monitoring objects (black fields – values above the detection limit, white fields of values below the detection limit, green period "background" data before the experiment, yellow period data during the experiment).
The presence of more or less similar PPCP substances in the river as in the outlet from the WWTP is not surprising; this is because, especially in times of low flows, inflows from treatment plants represent a nonnegligible part of the flow of watercourses in Europe. This fact is documented by comparing the contents of PPCP substances in some Central European (CE) rivers (see Tab. 2). The water quality in the Morava River in terms of the range of detected PPCP substances fits very well into this regional context; in absolute values the contents of some substances are even significantly higher (e.g. Oxypurinol, Iopromide, Telmisartan).
Table 2. PPCPs occurrence in six watercourses in CE region (The colour indicates the frequency of occurrence: red in all rivers; orange - the substance was missing in one of the streams; yellow - the substance was missing in two streams; olive - the substance was missing in three rivers.; Hrkal in Reberski, Selak 2022.)
Despite the similarity of PPCP substances, large quantitative differences occur between the outlet of WWTP and the Morava River. The highest average concentration in treated wastewater was exhibited by Oxypurinol (13 903 ng/l; Tab. 1). For the other eight substances, average content was higher than 1 g/l; these were Valsartan acid, Lamotrigine, Diclofenac, Telmisartan, Tramadol, Benzotriazole, Hydrochlorothiazide, and Sucralose. All these substances can also be found in river water, but in lower orders of magnitude – due to the dilution, degradation, or gradual transformation into others . However, such high differences in content are not always a general rule. In Table 1, the group of substances occur where the differences in PPCP content are smaller, or negligible, between the WWTP outlet and river Morava. These are, for example, 4-formylaminoantipyrine, Gabapentin, or Valsartan; the content of Metformin is more or less equal, with Acesulfame even more in the river water than in the WWTP discharge.
The data from Table 1 also shows the significant role of infiltrated water from the Morava River: the adjacent well PZ5 contained a significantly more varied range of PPCP substances before the start of the infiltration experiment than wells CP1 and PZ1 further from the river. The effect of the infiltration of treated wastewater in the VSK well was evidently not manifested in the PZ5 well.
4.3 Cluster analysis
Confirmation of changes in groundwater chemistry in wells PZ1 and CP1 due to the infiltration of treated groundwater into the VSK well was confirmed by statistical processing using cluster analysis. Cluster analysis is one of the methods that deal with investigating the similarity of multidimensional objects and classifying them into classes. It allows assessment of the relationships between individual clusters.
In the case of the implemented experiment, 30 chemical analyzes of water for PPCP substances were available, each of which contained a determination of the content of 113 substances (matrix of 30 x 113 elements). Cluster analysis focused on:
the number of clusters of PPCP substances – i.e. substances belonging to one set;
a measurable degree of difference in the occurrence of pharmaceuticals between individual observation objects.
The graphical output of this analysis is a dendrogram which allows measurable evaluation of the difference in the framework of the entire data matrix. The differences between groundwater, Morava river water, and infiltrated water during the whole experiment are documented in Fig. 4.
he results showed that the groundwater in wells PZ1, CP1, and PZ5 was significantly different from the output from the WWTP and mostly from the Morava River throughout the experiment. The water in Morava River is significantly more "related" to groundwater than water from the WWTP. As a result of infiltration into the VSK well, the difference between the infiltrated water from the WWTP and groundwater has demonstrably decreased. This statistically confirmed the change in groundwater quality due to infiltrated water originating from the WWTP.
4.4 Characterization of the transport process
The transport of PPCPs from the VSK infiltration well to the PZ1, CP1, or PZ5 boreholes can be analysed based on increases in the content of PPCP substances. Temporal changes in content enable the approximate assessment of transport rates. The most obvious markers are potentially those substances which did not occur in monitoring wells at the beginning of the experiment, but appeared later. Emphasis was placed on substances that, based on their physical and chemical properties, could have been considered tracers. Such non-degrading and non-sorbing substances migrate in the rock environment at the same speed as groundwater flows - i.e. without retardation. Oxypurinol and Benzotriazole were supposed to be in this category.
Table 3 Time development of Benzotriazole content at monitored objects
Table 4 Time development of Oxypurinol content at monitored objects
A comparison of both substances reveals that their content are significantly higher in treated wastewater than in the Morava river (Tab. 3 and Tab. 4). In well PZ5, the content of both pharmaceuticals were already present before the start of the infiltration experiment and remained more or less stable during the entire test, or there was no fundamental fluctuation. Neither Benzotriazole nor Oxypurinol was present in the CP1 and PZ1 boreholes before the start of the infiltration test. Both substances appeared in these objects with a delay of 8-12 days from the start of the experiment and had an upward trend. The values given correspond to velocities of 6.6 – 4.4 m/day. The distance between VSK and CP1 wells is 52.5 m.
4.5 Generalization of PPCP substance behaviour
The results of the infiltration experiment at the Kojetín site enable PPCP substances to be divided into four groups
A/ Substances with random occurrence in very low concentrations and therefore with an undetectable origin
At the WWTP outlet, the substances of this group often had concentrations below the detection limit, so the infiltration experiment could not affect the concentrations of these substances in the monitoring wells PZ1, CP1 and PZ5. Nevertheless, they were analyzed in low concentrations at irregular intervals in some samples. An example of this group of substances is caffeine, which was detected in well PZ1 in the sample from 11/20 and in well CP1 in the sample from 11/30. However, during the entire experiment, caffeine was not detected in the WWTP water samples. A similar result can be observed with other substances such as Bisphenol S, Methylparaben, Paracetamol, Paraxanthine, Propylparaben, and Saccharin.
B/ Substances with a proven negative impact on the Quaternary aquifer
Substances of this group are characterized by a high concentration in treated wastewater, which was later manifested in the occurrence in wells PZ1 and CP1. It is important that these substances were not present in the wells before the start of the experiment. The resulting content of these substances in the wells remain lower than at the WWTP outlet. Characteristic representatives of this group are oxypurinol, Sucralose, Hydrochlorothiazide, Diclofenac-4-hydroxy, and Benzotriazole. But all the substances which systematically occurred during the infiltration experiment should be included (Tab. 2). Absence of these substances at the beginning of the experiment reveals that they are degrading at some rate in the groundwater.
C/ Substances persistent, apparently without significant WWTP influence
Substances in group C are characterized by a higher concentration in wells than in treated wastewater. They usually occurred in the groundwater from the beginning of the infiltration experiment. The water in the observed wells was therefore closer to water from the Morava River than to the WWTP. A typical representative is Acesulfame, whose concentrations in the river exceeded twice the average values of the water leaving the WWTP. This group also includes DEET and Sulfamethazine, which are all persistent substances that are not subject to degradation either in the Morava River or in the rock environment of the gravel sands of the experimental site.
D/ Substances with a high natural attenuation capacity.
Substances in group D are characterized by concentrations above the detection limit in the water leaving the WWTP, possibly also in the Morava River. Nevertheless, these substances were not detected in the observation wells. It can be assumed that these substances have a good sorption/degradation capability in the rock environment. This is a relatively large group of substances that includes Acebutolol, Atenolol, Azithromycin, Benzotriazole 1-methyl, Bisoprolol, Celiprolol, Cetirizine, Citalopram, Clarithromycin, Climbazole, Clindamycin, Cyclophosphamide, Erythromycin, Fexofenadine, Fluconazole, Irbesartan, Carbamazepine-2- hydr., Carbamazepine-DHH, Carbamazepine-E, Lamotrigine, Iosartan, Metoprolol, Mirtazapine, Naproxen, Naproxen-O-desmeth., Oxcarbazepine, PFOS, Sertraline, Sitagliptin, Sotalol, Sulfamethoxazole, Telmisartan, Tramadol, Trimethoprim, Venlafaxine, Venlafaxine O-desmet, Verapamil, and Warfarin.