Conference Agenda

The deployment of IoT platforms provides accurate data acquisition and enhances algorithms for remote plant monitoring.

The proposed IoT node enables automated irrigation and data-driven decision support by acquiring sensor data related to soil, environment, and plant health conditions. The resulting IoT network is highly scalable, supporting variable numbers of nodes and sensors per node.

The measured soil parameters are temperature, moisture, and matric potential. Environmental parameters include air temperature, pressure, humidity, and water flow within irrigation pipes. Plant health status is assessed through a custom sensor measuring the bioimpedance of the stem, and it is integrated with leaf transpiration measurements.

The network is LoRaWAN-based, thus ensuring low-power, long-range wireless communication. Moreover, it leverages existing IoT infrastructures and compact, low-cost, and maintainable nodes, thereby minimizing capital investment even in large-scale deployments. An idle current below 10 μA enables an estimated battery lifespan exceeding 10 years (measurement routines every 30 minutes; SF7 LoRaWAN transmission).

The resulting IoT platform can maintain or enhance crop productivity, while preserving water resources through prediction of water-deficit stress and optimization of irrigation scheduling in terms of timing and water volume. The system is adaptable to various agricultural contexts, including fields, orchards, greenhouses, and research facilities.

Efficient agricultural water management is critical in arid regions such as Iran, where over 90% of water consumption is attributed to agriculture. This study presents a 24/7 operational workflow for the daily simulation of key biophysical and agrometeorological variables, including evapotranspiration (ET), biomass production, NDVI, and Albedo, covering Iran as the simulation domain. The system integrates three subsystems: the WRF regional atmospheric model for operational weather forecasting, VIIRS and Sentinel-2 satellite data retrieval and preprocessing, and the SEBAL surface energy balance model implemented via pySEBAL. Products are generated at two spatial resolutions of 375 m and 100 m. Verification of reference ET against observations from national meteorological stations across four low-cloudiness days yielded RMSE values generally within 2 mm/day, demonstrating reasonable model performance at the country scale. Time series analysis over sugarcane fields confirmed realistic seasonal dynamics in ET and biomass. The system operates on a High Performance Computing infrastructure and delivers outputs through a web-based geospatial service for use by farmers and water resource managers.

Excessive total phosphorus (TP) export from agricultural land is a major cause of eutrophication in Lake Erie. This study uses a large-scale SWAT model to evaluate TP reduction across the Canadian Lake Erie Basin (21,750 km²). The model was calibrated and validated for streamflow, sediment, and TP, yielding satisfactory results, and was used as a baseline scenario. The analysis focuses on how Critical Source Area (CSA) identification and BMP combinations affect TP reduction in five major tributaries at HRU and outlet levels: Grand, Thames, Sydenham, Big Otter Creek, and Big Creek. Results indicate that combined BMPs (fertilizer reduction plus cover crops) achieved much larger reductions than individual BMPs. The TP reduction increased nonlinearly with the implementation area. At full HRU coverage, combined BMPs reduced TP loads by about 63–67% in Big Otter Creek, 55–58% in the Grand River, 52–55% in Big Creek, 40–42% in Sydenham, and 45–48% in Thames. Overall, the watershed-scale performance of BMPs depends on how they are targeted and combined, as well as on hydrological connectivity. These findings highlight that context-specific CSA identification and the implementation of bundled BMPs are essential to meet phosphorus reduction targets.

Despite the importance of plastics in agriculture, their intensification generates significant environmental issues, connected to agricultural plastic waste (APW) management and soil pollution risks. Current remote sensing typically provides static mapping, lacking spatially explicit management support. This study develops an integrated remote sensing and GIS framework for mapping agricultural plastics (AP) and assessing spatial risk in the Basilicata region (Southern Italy). Using multi-temporal Sentinel-2 imagery within GoogleEarth Engine, the Adaptive Plastic Mulch Index (APMI) and Advanced Plastic Greenhouse Index (APGI) have been implemented to differentiate plastic types - including greenhouses, black/white mulch films, protective nets, etc. AP maps have been converted into APW estimates within a GIS environment, using crop‑ and use‑specific plastic coefficients. A composite Agricultural Plastic Pollution Risk Index (APPRI) has been then computed, by multiplying APW estimates by the relative risk indices (RRIs). To automate prioritization, a dynamic degradation-risk proxy, based on cumulative UV radiation and surface temperature, was integrated to predict potential fragmentation in the study area. Applied in the Metapontino district (Basilicata region), this framework contributed to identifying high-risk hotspots for collection prioritization. This scalable tool enables informed territorial planning and circular resource management, triggering interventions before material degradation occurs, thereby mitigating microplastic soil contamination.

Mediterranean regions are particularly vulnerable to erosion due to different factors. Forest fires increase the risk of soil erosion by reducing vegetation cover and altering soil properties, such as infiltration capacity. Furthermore, climate change is expected to intensify soil erosion processes in areas affected by burning. Soil bioengineering represents a low-impact Nature-Based Solutions (NBSs) that can help counteract post-fire erosion.

The study aims to analyse the effects of soil bioengineering techniques in the pilot area of Gravina in Puglia, covering 1,900 hectares, where a fire affected 1,170 hectares on 12 August 2017.

The works were carried out in two experimental sites located on slopes with similar characteristics—such as gradient, soil type, and fire severity—but different exposure. The works included weed removal, installation of wattles and palisades, and planting of native shrubs and tree species. A monitoring plan combining field surveys and remote sensing activities was developed to evaluate the effectiveness of the soil bioengineering works.

After two years, the first site showed a high survival rate of the planted species, while in the second site spontaneous plants had a very intense development. The growth of the vegetation, detected using NDVI values, showed a significant increase in both areas.

Electromagnetic induction (EMI) sensing is increasingly employed for soil zoning in precision agriculture, with recent advances moving beyond merely mapping apparent electrical conductivity (ECa) to more physically based, process-oriented zoning frameworks. These methods incorporate transparency at the signal level, control over data acquisition architecture, and agronomic interpretability within multi-layer spatial analysis. An open-source Raspberry Pi–based data acquisition (DAQ) system was developed through reverse engineering of the EMI sensor’s analogue output architecture, enabling direct extraction, conditioning, and digitisation of the EM38 (Geonics Limited) output while overcoming the limitations of proprietary closed systems. This architecture guarantees full signal traceability, configurability, and modular integration within digital agriculture workflows, while maintaining functional equivalence with traditional platforms. Methodological refinements further addressed structural and hydraulic factors influencing spatial variability. Post-tillage soil surface roughness and aggregate rearrangement were found to affect ECa patterns, and in tree orchards with non-uniform irrigation, ECa may also reflect moisture redistribution caused by irrigation alongside inherent soil heterogeneity. Under these circumstances, single-layer EMI mapping may give biased zoning results. Combining ECa with vegetation vigour indicators derived from NDVI improved the separation of soil structural variability from plant water response, reducing confounding effects, and resulting in more stable clustering.