Hydrogeological and multi-isotopic approach to define nitrate pollution and denitrification processes in a coastal aquifer (Sardinia, Italy)

Agricultural coastal areas are frequently affected by the superimposition of various processes, with a combination of anthropogenic and natural sources, which degrade groundwater quality. In the coastal multi-aquifer system of Arborea (Italy)—a reclaimed morass area identified as a nitrate vulnerable zone, according to Nitrate Directive 91/676/EEC—intensive agricultural and livestock activities contribute to substantial nitrate contamination. For this reason, the area can be considered a bench test for tuning an appropriate methodology aiming to trace the nitrate contamination in different conditions. An approach combining environmental isotopes, water quality and hydrogeological indicators was therefore used to understand the origins and attenuation mechanisms of nitrate pollution and to define the relationship between contaminant and groundwater flow dynamics through the multi-aquifer characterized by sandy (SHU), alluvial (AHU), and volcanic hydrogeological (VHU) units. Various groundwater chemical pathways were consistent with both different nitrogen sources and groundwater dynamics. Isotope composition suggests a mixed source for nitrate (organic and synthetic fertilizer), especially for the AHU and SHU groundwater. Moreover, marked heterotrophic denitrification and sulfate reduction processes were detected; although, for the contamination related to synthetic fertilizer, the attenuation was inefficient at removing NO3− to less than the human consumption threshold of 50 mg/L. Various factors contributed to control the distribution of the redox processes, such as the availability of carbon sources (organic fertilizer and the presence of lagoon-deposited aquitards), well depth, and groundwater flow paths. The characterization of these processes supports water-resource management plans, future actions, and regulations, particularly in nitrate vulnerable zones.

The forage cropping systems are based on double-cropping silage corn -Italian ryegrass (representing > 134 80% of the irrigated land), and ~35,600 dairy livestock are raised in a narrow area (Giola et al. 2012).

136
The climate is Mediterranean, and mean annual temperature and precipitation are 16.7 °C and 568 mm, 137 respectively . Some 73% of annual rainfall occurs between October and March (Demurtas et al. 138 2016), and the average annual aridity index (rainfall / reference evapotranspiration) is 0.49 (semiarid area).

140
The study area occupies the northern part of the Campidano rift, with a landscape characterized by

237
The AHU is recharged laterally from the VHU and by zenithal infiltration from both precipitation and irrigation

281
Brackish-salt-water occurrences are limited to a few wells on the coast (SHU) and inland in the AHU.

282
These waters showed a Ca/HCO3 ratio > 1. However, waters on the coast had high chloride and were 283 oversaturated with respect to calcite, whereas inland waters showed SO4 up to 340 mg/L and 284 undersaturation with respect to calcite.
Brackish waters were mainly found in the SHU, in the southern part of the reclamation area. Chemical 286 compositions range from NaCl to CaCl types, with very variable nitrate content (L, M, HH) and high HCO3.

301
According with the water isotope values of groundwater, no significant influence of seawater intrusion is 302 observed. Also the relationship between chlorine content and water isotopic values points out that seawater 303 intrusion is not occurring in the studied area. Some evaporative processes in the unsaturated zone can 304 explain the displacement of some data from the water meteoric line to the global meteoric line.

440
The relationship between δ 13 CDIC and bicarbonate concentrations, as well as NPDOC concentration, is 441 typically used to explain the role of organic matter oxidation in denitrification. In the study area, measured

497
In the SHU, the variation in NO3content is related to the heterogeneous nitrate load and the degree of 498 denitrification, the latter being modulated by well depth, the availability of organic C, and the groundwater 499 flow path. In general, a lower content of nitrate in groundwater, imputable to denitrification processes, has been observed when nitrate comes from organic rather than synthetic fertilizers. It is reasonable to assume 501 that the high NO3load from synthetic fertilizer overwhelms the capacity to promote efficient denitrification.

502
This is in line with the situation in the study area over recent decades, in which organic effluents 503 (representing > 50 % of the crop N input source) have been progressively replaced by mineral fertilizer not 504 reducing, de facto, nitrate leaching.

505
The wells with depths < 10 m b.g. had higher nitrate concentrations, especially when the origin was clearly 506 attributable to synthetic fertilizers ( Figure 12). In addition, those wells showed evidence of sulfate reduction 507 ( Figure 13).

554
In the SHU west of the reclaimed lagoon, a mixed source for nitrate was defined, which in some cases 555 increased according to the flow path. Also in this groundwater, there were heterotrophic denitrification and 556 sulfate reduction processes. However, high nitrate concentrations in samples influenced by synthetic 557 fertilizer indicate attenuation insufficient to remove NO3to below the human consumption threshold. For 558 SHU groundwater, therefore, the change from organic to inorganic fertilizer application did not reduce the 559 nitrate pollution in groundwater. As in the study area natural attenuation processes are taking place, it would 560 be suitable to promote biostimulation to add dissolved organic matter and, consequently, electron donor 561 availability to the aquifer therefore promoting denitrification.

562
In conclusion, the present study demonstrates that the combination of multi-isotope, chemical and 563 hydrogeological approaches facilitates nitrate source identification and the evaluation of nitrate removal via 564 biogeochemical transformation processes in a multi-aquifer system under Mediterranean conditions.

565
However, characterization of potential nitrate and marine source terms as well as the enrichment factor for 566 aquifers should be acquired to better define the evolution of contamination and to implement remediation