The Eastern Mediterranean region, including Cyprus, has experienced warming at twice the global average rate, exceeding the thresholds set by the Paris Agreement. Cyprus experiences some of the most severe freshwater shortages in the EU, significantly affecting both agricultural productivity and ecosystem health. The frequent depletion of surface water resources has led to the overextraction of groundwater, already placing the majority of aquifers under serious threat (Sofroniou and Bishop, 2014). This issue is especially severe in Kokkinochoria, Cyprus’s main potato-producing region, where excessive groundwater pumping, mainly to meet the water demands of potato cultivation, has led to seawater intrusion and nitrate contamination (National Open Data Portal Cyprus, 2024). The problem arises from the fact that, although potatoes are highly sensitive to water stress and nutrient deficiencies (Wang et al., 2021), only 40%–60% of the nitrogen applied is actually absorbed by the plants. The remainder is often leached away through excessive irrigation, posing significant environmental risks (Jones et al., 2021). To overcome this problem, regulatory measures like the Nitrates Directive (1991) within the Water Framework Directive (2000) have been implemented, yet improving resource efficiency remains vital. In recent years, IoT-based sensor practices are being promoted to optimize fertigation in potato systems, reducing nitrate and phosphorus leaching (Berbel and Martínez-Dalmau, 2021). Despite available data and smart farming tools (e.g., Incrocci et al., 2020; Adamides et al., 2020), implementation across crops and regions is complex due to variability in soil and plant characteristics. This complexity limits the adoption rate in small-scale farming (Huber et al., 2023), highlighting the need for tailored, region-specific, and species-specific smart farming solutions.

The potato, the most widely grown non-grain food crop, combines high yield and nutritional value, making it vital for global food security (Economou et al., 2023). In Cyprus, potato is a key crop with a utilized agricultural area of 4,000 ha, producing 110,000–120,000 tons annually (70%–75% of which is exported), while being produced under irrigation consuming over 12 million m³ of freshwater. Planting occurs from November to early February, with harvest from March to mid-June. Potatoes are highly sensitive to soil moisture and nutrient deficits, particularly during tuber development, due to a shallow rooting system. To ensure yields, growers often overapply water and nitrogen, leading to high nitrate (NO₃ ⁻) leaching risks (Venterea et al., 2011; Giletto et al., 2019). Excessive inputs also increase nitrogen losses through leaching, denitrification, and volatilization, with leaching accounting for ~85% of total N losses under irrigated conditions (Zebarth and Rosen, 2007). Despite the importance of N₂O and NH₃ emissions from fertilized crops, their emissions from potato fields in Cyprus have never been measured. Elsewhere, N₂O losses in irrigated systems range from 0.21 to 3.06 kg N ha⁻¹ (Huang et al., 2024). Phosphorus (P) losses via runoff and leaching, due to soil saturation, further exacerbate eutrophication risks (Economou et al., 2023). Therefore, as one of Cyprus’s most water-intensive crops, even minimal irrigation savings can yield significant environmental benefits (Christou et al., 2017; Dalias et al., 2018). Enhancing ferti-irrigation in spring-grown potatoes in Cyprus with sensor-based technology can improve water use efficiency and reduce nutrient losses (Steyn et al., 2016). This approach supports the European Green Deal and aligns with EUMED9 (2024) calls for innovative solutions to mitigate water scarcity impacts on regional crop production.

Since water management is closely linked to nutrient runoff, deploying IoT-based sensor farming strategies offers clear environmental benefits compared with conventional practices (Steyn et al., 2016). However, the effectiveness of such smart systems varies with factors like climate, cultivar choice, and harvest timing, highlighting the need for further research and much improvement in scientific literature. Based on this, the research hypothesis was established: the gaiasense™ system enhances resource-use efficiency and reduces nutrient losses while maintaining the yield and quality of spring potatoes under Mediterranean water-scarce conditions. To test this, the study evaluated whether a regionally calibrated IoT-connected sensor-based ferti-irrigation system (gaiasense™) could enhance water productivity and nutrient use efficiency, reduce nitrogen and phosphorus losses, and sustain tuber yield and quality in spring potato production in Kokkinochoria, Cyprus, compared with conventional farming practices.

Efficient irrigation management is pivotal for the sustainability of irrigated agriculture, particularly in semiarid regions like Cyprus. The GS (gaiasense™) treatment applied 499.4 mm of irrigation, representing a 31.5% reduction compared with the 728.7 mm applied in the conventional CL treatment. Most of the savings occurred during the later stages of growth, with a maximum reduction of 40.2% ( Table 1 ). As a result, total drainage losses were markedly lower under GS (43.5 mm) than CL (272.8 mm, Table 1 ), indicating significantly higher water excess and mismanagement in the CL regime. This suggests that CL frequently exceeded crop water needs, increasing the risk of irrigation-induced leaching and inefficient use of water ( Table 1 ; Venterea et al., 2011). Slightly negative drainage values of the GS treatment may reflect more precise water matching with crop demand or even an upward capillary movement. In any case, a soil water content at the upper soil layer within 25–35% (per volume of water content) should be maintained during the whole potato cropping period (Adamides et al., 2020; Figure 1 ). This suggests that the GS regime more closely aligned irrigation supply with crop water demand. As a result, GS succeeded a notably higher water productivity of 12.66 kg/m³, compared with 9.25 kg/m³ under CL (an increase of 37%). These results are consistent with values reported in studies from Europe and the Mediterranean region, which typically range between 3 and 15 kg/m³ (Steyn et al., 2016). Thus, the results clearly demonstrate that optimized irrigation scheduling not only reduces water waste but also enhances production efficiency per unit of water applied. Excessive irrigation often leads to substantial drainage, which in turn increases the risk of nitrate (NO₃ ⁻) leaching, as NO₃ ⁻ readily moves with drainage water. In irrigated potato systems, nitrate leaching is the primary pathway of nitrogen (N) loss, accounting for approximately 85% of total N losses (Zebarth and Rosen, 2007), and this process seriously threatens groundwater quality (Thompson et al., 2007). Indeed, preliminary measurements of nitrogen (N) gas emissions in the current showed negligible N₂O emissions (<0.2% of applied N), with ammonia (NH₃) as the primary gaseous N loss. Phosphorus losses from P-saturated soils via leaching and runoff strongly contribute to freshwater eutrophication (Giletto et al., 2022). In this study, reduced P application effectively lowered phosphate concentrations in runoff and leachate ( Figure 2 ). These losses, caused by rainfall or excess irrigation, would have contributed with nitrate to surface water eutrophication (Giletto et al., 2022). Thus, improving irrigation control by shifting from regional averages of evaporation and rainfall to site-specific microclimatic measurements ( Figure 1 ) using IoT-based weather stations is a strategic approach to reduce water and nutrient losses. This study reveals that sensor-based systems optimize irrigation in potato production, conserving water and supporting yield, sustainability, and climate adaptation in the Eastern Mediterranean (Steyn et al., 2016).

Mismanaged fertilizer application increases nutrient losses and environmental degradation (Economou et al., 2023). Here, fertilizer timing was consistent in both treatments, following a split-application strategy ( Table 1 ) aligning with best practices (Gianquinto et al., 2013; Sha et al., 2022). Fertilizer doses ( Table 1 ) under the GS regime were substantially reduced compared with CL, with total applications of N, P, and K decreased by 9.1%, 27.9%, and 53.6%, respectively. Despite lower macronutrient inputs, GS featured a targeted increase in micronutrients, particularly Zn, Mn, and Cu, likely to support crop performance. Both treatments received relatively high N inputs ( Figure 3B ) to accommodate high yields exceeding the national average (35.6 t/ha, Economou et al., 2023) and this probably diminished yield effects between treatments ( Figure 3A ). Indeed, the crop’s good nutritional status was evidenced by comparable N and P uptake across treatments ( Figure 2 ), high yields (52–59 t/ha), uniform large tubers (avg. 225 g), high dry matter (19.5%), and low nitrate content (68 mg NO₃/kg FW), meeting European market standards. Similar production levels and nitrogen applications in both treatments resulted in no significant differences in tuber yield per unit of nitrogen, reflecting comparable nitrogen use efficiency ( Figure 3C ). Total P inputs were significantly lower in GS without reducing yield ( Figures 3A, B ). GS also increased tuber yield per unit of P ( Figure 3D ) and markedly reduced potential N and P losses by over 50% compared with CL ( Figures 3E, F ). As a result, GS showed a more favorable yield to loss ratio, producing over twice as much tuber biomass per unit of N and P lost ( Figures 3G, H ) This pattern was confirmed by nutrient (N, P) mass balance analysis, highlighting the agronomic and environmental outcomes of the sensor-based GS strategy ( Figure 2 ). Particularly, nitrogen calculations showed that GS treatment resulted in significantly lower nitrogen losses (51.5 kg N/ha) than CL treatment (129.4 kg N/ha, Figure 3E ), mainly due to higher residual soil nitrogen and slightly lower fertilizer input ( Figure 2 ). Phosphorus calculations revealed higher P losses in CL treatment (78.0 kg P/ha) compared with GS treatment (37.4 kg P/ha, Figure 3F ), in agreement with a greater fertilizer input (121.8 vs. 87.2 kg P/ha) and slightly lower residual soil P ( Figure 2 ). Noteworthily, residual soil nitrate at harvest was markedly higher in GS compared with CL ( Figure 2 ), indicating improved soil fertility and a greater potential for nitrogen carryover effect. This can be explained by the fact that the impact of drainage water was minimal during the later stages of potato growth in GS, indicating limited or no excess water percolation beyond the root zone. Although residual N can be a potential loss pathway (Clément et al., 2020; Bohman et al., 2021), residual nitrate–nitrogen up to certain levels may remain relatively stable within the soil profile. Under conditions unfavorable to leaching, such as the clayey red soils and dry summer conditions observed here, this may result in minimal leaching risk. Therefore, a rotational crop planted after potatoes and before the onset of the rainy season could be used to recover residual soil nitrogen (Bohman et al., 2021; Wu et al., 2022). Lower N application rates might achieve similar yields with less residual NO₃ ⁻–N at harvest, but this requires further study. These results confirm that the GS strategy conserves water and nutrients and enhances their efficiency, supporting sustainable potato production in Mediterranean conditions.

A reduction exceeding 50% in potential nitrogen and phosphorus losses was achieved under the GS (sensor-based) treatment compared with the conventional CL system, alongside a 37% increase in water productivity as mentioned elsewhere. Furthermore, the dataset reveals a direct correlation between increased drainage volumes and higher N and P losses under CL, supporting findings by Giletto et al. (2022), who reported a strong link between drainage and nutrient leaching. This suggests that substantial N and P losses occurred beyond the root zone under conventional management, raising environmental concerns. Similarly, Economou et al. (2023) reported that nutrient runoff, particularly N, P, and SO₃, significantly contributes to freshwater eutrophication in spring potato systems in Cyprus (2.19 kg P eq./t). Interestingly, while agronomic efficiency (kg tubers per kg N supplied) was similar across treatments, the yield to loss ratio (kg tubers per kg N lost) more than doubled under GS ( Figure 3G ). A similar pattern occurred for P, where the yield to P loss ratio far exceeded yield to P supply ratio ( Figures 3D, H ), reflecting the environmental rather than agronomic perspective. However, achieving a better synchrony between the two is fundamental in modern intensive agriculture and it demonstrates that sustainability does not necessarily come at the expense of productivity (Souza et al., 2021). In this direction, the IoT sensor-based tools can save water and minimize nutrient leaching, without compromising productivity in potato cropping systems, aligning technological improvements in the fields of nutrients use, irrigation, and sustainability (Berbel and Martínez-Dalmau, 2021). Indeed, compared with conventional farming (CL), the GS (gaiasense™) system reduced variable and total costs by 25% and 19%, respectively, with notable decreases in fertilizer (37%), pesticide (51%), irrigation (32%), and labor (10%) costs. Gross profit and irrigation water productivity increased by 24% and 22%, respectively, whereas no pesticide residues exceeded maximum limits (data not shown). These benefits are especially relevant to small-scale farming and to climate hotspot regions facing water scarcity and degraded aquifers, such as the Eastern Mediterranean, empowering environmental sustainability in potato farming systems, suggesting while also highlighting that socioeconomic and environmental sustainability are not mutually exclusive (Stylianou et al., 2020). Nevertheless, this study was limited to a single growing season in Kokkinochoria, Cyprus, under specific soil and climate conditions, which may affect the generalizability of results to other regions or years. Moreover, while the system was calibrated for potatoes, its application to other crops would require crop-specific and site-specific adjustments. Multiyear, multicrop trials across diverse agroecological zones may be needed to confirm the broader applicability of these findings.

This study highlights the critical role of IoT-based sensor ferti-irrigation strategies in enhancing environmental sustainability and resources (water/nutrient) efficiency in intensive potato cropping systems in Cyprus. By integrating IoT-connected microclimatic monitoring into irrigation management, the GS system achieved environmental gains without sacrificing tuber yield or quality. Further research is ongoing to determine whether reduced nitrogen inputs can sustain these outcomes in high-yield potato crops, particularly when implementing the system on a larger scale and across diverse growing conditions. Overall, this approach exemplifies how smart technologies can bridge the gap between agricultural productivity and environmental protection, advancing the objectives of the EU Green Deal contributing to climate-resilient farming systems.