Conference Agenda

This work presents Smartcrop DSS, a decision support system for fertilization and crop management based on Food and Agriculture Organization (FAO) agronomists’ workflows in Caribbean regions. The system is implemented as a web application integrated with a RESTful API and database. Smartcrop DSS determines optimal fertilization programs using soil analysis, production system (open field or hydroponic), target yield, crop nutrient requirements, and a user-defined set of fertilizers. It applies a linear programming model derived from FAO spreadsheet tools to generate a fertilization calendar with daily and total costs, minimizing fertilizer inputs while satisfying crop nutrient constraints. For hydroponics, the system uses an iterative ionic balance algorithm that combines key ion vectors with pH acid neutralization to ensure nutrient solution stability. An additional module identifies suitable pesticides from a product database based on crop symptoms or diseases, including FRAC-based filtering. Currently in test release, the system is being evaluated by agronomists for effectiveness and usability. Its adoption aims to reduce fertilizer use, costs, and environmental impacts while supporting policy monitoring. The project is designed for future expansion to other Caribbean and South American countries.

Several irrigation decision support systems (DSS) have been developed to improve water management. However, many rely on rigid, predefined data workflows tailored to specific sensor types or models. This rigidity reduces DSS portability and usability and can weaken accuracy when inputs are incomplete and hydrological conditions are non-standard. This research addresses this gap by developing an adaptive framework that links DSS complexity to the information available on soil–plant–atmosphere processes and to site-specific constraints. The work comprises: a state-of-the-art analysis of how crop water requirements and soil water status are represented in model-based and data-driven approaches; a modular workflow to harmonise inputs and explicitly account for uncertainty when switching among alternative parameterisations and data sources; and a field assessment in a post-flood alluvial environment (Argenta, Northern Italy) used as a case to stress-test transferability. The model-based DSS is tested under alternative input configurations based on estimated versus measured soil, groundwater, and meteorological information. In parallel, a non-invasive data-driven approach using multi-source indicators (Cosmic-Ray-Neutron-Sensing, NDVI and Vapor-Pressure -Deficit) is evaluated in the same context. The expected outcome is a transparent basis for selecting, within one framework, the least complex DSS configuration that remains usable and reliable for a given scenario.

Water scarcity is among the most critical challenges facing Mediterranean agriculture, particularly rice production, which requires substantial amounts of water. Addressing this challenge requires the adoption of innovative and sustainable tools for water management systems. The objective of this study was to co-design a water monitoring and management system for rice cultivation under Mediterranean conditions. The approach was based on a multi-stakeholder participatory framework involving farmers, farm advisors, water management authorities, researchers, and other relevant actors to ensure that the developed solutions respond to local needs and practical constraints. The system integrated smart agriculture technologies, including field sensors for monitoring soil and water conditions, as well as advanced remote monitoring tools such as Unmanned Aerial Vehicles (UAVs). These technologies were combined with field-based measurements to support data-driven decision-making and improve irrigation efficiency. The study found that the tools used optimized water use while maintaining crop productivity and minimizing environmental impacts. By integrating digital technologies with participatory design processes, the study demonstrates how smart farming can support sustainable rice production, enhance resource efficiency, strengthen food security, and increase resilience to climate variability and future environmental changes.

Thermal indices used for irrigation scheduling require calibration under local climatic conditions. In Uruguay, characterized by high humidity and frequent cloud cover, their applicability remains largely untested. This study aimed to calibrate and evaluate thermal indices for maize under different irrigation strategies and establish CWSI baselines using lysimeters. The experiment was conducted at INIA Las Brujas, Uruguay, using 12 drainage lysimeters under a rain-out shelter with a high-yield maize hybrid (120,000 plants ha⁻¹). Four irrigation treatments were evaluated: 100%, 80%, and 50% ETc replacement, and a threshold treatment (irrigation restored to field capacity when the FAO depletion threshold was reached, p = 50% TAW), with three replications. Actual evapotranspiration was measured in each lysimeter plot, and continuous canopy temperature was recorded using nadir-pointing infrared sensors. Midday stomatal conductance, crop cover, and weekly NDVI were measured. Soil water content was recorded three times per week at 0.15, 0.30, 0.60, and 0.80 m depth using a neutron probe. Thermal indices clearly differentiated irrigation treatments. CWSI baselines were established locally, and canopy temperature correlated with stomatal conductance and soil water content. NDVI detected sustained stress but was less sensitive to early stress. Thermal indices show strong potential for irrigation scheduling.

A decision support system (DSS) for viticulture irrigation was presented to enhance water balance amid increased climate variability. The methodological framework incorporates agro-hydrological modeling, multispectral remote sensing, and soil pedophysical characterization for tailored and data-informed irrigation management. The FAO-56 Penman–Monteith equation was used to figure out reference evapotranspiration (ETo), and the dynamic crop coefficient (Kc) was used to figure out crop evapotranspiration (ETc). Kc was based on real-world relationships with NDVI. The NDMI index, acquired from Sentinel-2 images in Google Earth Engine, was used to evaluate the status of grapevines. Geophysical surveys were used to find out the available water capacity (AWC) and to describe how the hydraulic properties of soil change over space. This data goes into a daily water balance model that is part of an interactive digital platform. The system architecture is based on open-source tools and datasets. This makes sure that the methods are clear, the science can be repeated, and the operations can grow. It also makes it easier to add machine learning algorithms later to improve adaptive predictive models. The results show improved spatial and temporal representativeness of irrigation estimates, leading to reduced distributed volumes and increased water use efficiency (WUE).

Water scarcity is a persistent challenge for Mediterranean agriculture, intensified by climate change. Irrigation is often managed using fixed historical averages, ignoring interannual variability that strongly affects crop water demand, leading to inefficient allocation and yield losses. Seasonal Weather Forecasts (SWFs) provide probabilistic information months in advance and can support strategic crop and irrigation planning. This study assesses the applicability of ECMWF SEAS5 seasonal forecasts to estimate maize development and irrigation requirements in continental Portugal. Daily ensemble forecasts with a 7-month lead time (April–October) were analysed for 1987–2024, using 1987–2016 as baseline and 2017–2024 for evaluation. Bias-corrected AgERA5 reanalysis served as reference. Results show that SWFs consistently capture seasonal temperature and precipitation anomalies, although magnitude errors persist. Despite considerable ensemble spread in precipitation, forecasts successfully indicated drier or wetter seasons. Temperature forecasts were particularly effective in estimating crop cycle length using growing degree days, with maize phenology errors of about one week. A simplified soil water balance model showed tendency for underestimation of irrigation needs, mainly due to an underestimation of reference evapotranspiration. Overall, SWFs provide actionable early-season information, replacing static climatological assumptions with climate-informed scenarios to enhance adaptive water management under increasing variability.

The rewetting of peatlands aims to reduce overall CO₂ emissions; however, it may substantially affect grassland management. This study evaluated the trafficability of a rewetted peatland pasture during one season. Therefore, shear strength (SS) and volumetric soil moisture (vSM) were measured in the topsoil at the end of each grazing period with up to six beef cattle on four plots (P1: 0.88 ha, P2: 0.52 ha, P3: 1.00 ha, P4: 0.78 ha). Measurements were taken at ten predefined diagonal sampling spots (11 on P1). The mean duration of a grazing period ranged from 3.8±0.4 to 9.5±2.3 days. The number of grazing periods ranged from five to six. Plot had a significant effect (P<0.05) on both SS and vSM. Mean SS values were 25.15±2.91, 17.04±2.24, 32.60±3.62, and 48.61±10.36 kPa for P1, P2, P3, and P4, respectively. Mean vSM values were 79.83±2.52, 84.11±0.85, 75.60±5.51, and 72.66±7.86 % for P1, P2, P3, and P4, respectively. The measured values were consistent with visual observations of turf damage and cattle sinking while walking, particularly in some areas on P1 and P2. These findings underline the importance of adapting grazing management on rewetted peatland pastures while considering potential risks for people and animals.

Uruguayan citrus production targets high-quality fresh fruit export markets. Full netting systems (anti-insect and anti-hail) are increasingly adopted to mitigate climatic risks and reduce pest pressure. However, their impact on orchard microclimate and crop water use under local conditions remains insufficiently quantified. This study evaluated the effects of permanent netting on microclimate, plant physiological performance, and irrigation water productivity in commercial citrus orchards. The experiment was conducted in a 4-ha ‘Afourer’ mandarin (Citrus reticulata Blanco) orchard fully covered with 40-mesh or 9-mesh anti-bee netting, compared with a 2-ha uncovered control. Irrigation water use was quantified in each environment and water productivity was calculated and related to physiological variables. Leaf gas exchange (CO₂ assimilation, stomatal conductance, and transpiration) and chlorophyll fluorescence (Φ_PSII) were measured during summer. Net-covered environments reduced irrigation requirements by 24%, while yield increased, resulting in higher water productivity (45–100%). Netting increased maximum relative humidity by 10–15% and maximum air temperature by up to 3 °C. Plants under netting exhibited greater CO₂ assimilation, stomatal conductance, and transpiration during periods of high radiation, together with higher Φ_PSII values. These results indicate that permanent netting modifies orchard microclimate and improves physiological performance while reducing irrigation requirements.

Lawn areas are important in urban environments for well-being, leisure activities, play and sport. Maintaining these lawn areas by mowing them regularly has a significant impact on lawn quality.
Hence, the aim of our research was to develop methods for automatically assessing the quality of lawn mowing. The so-called leaf cut index is commonly used for this purpose. It represents the ratio of the grass cut edge produced by the lawn mower to a minimally invasive cut edge. Until now, individual grasses have been examined under a microscope and assessed manually.
We have developed a system that captures grasses in vivo using a special camera and determines the leaf area indices of individual grasses using computer-assisted image analysis. Individual grasses are selected using pattern recognition. The optimal cutting edge length is simulated by a theoretical cut perpendicular to the stem orientation. The actual cutting edge length is then determined and compared with the optimal cutting edge length. We have implemented the algorithm as a prototype and validated it using manual index determination.
Results comparable to manual assessment were achieved, but with minimal effort and much higher reproducibility. The system could be installed in lawnmowers to continuously evaluate the cutting.

Filtration in drip irrigation prevents emitter clogging through two alternating modes: filtration, which traps particles, and backwashing, which cleans the clogged filter. Filtration cycles follow different patterns that affect both the energy and water use of pressurized filters. This study develops a general framework to improve filter design by classifying pressure‑drop behavior during filtration as linear, piecewise linear, or quadratic, using data from disc, screen, and sand filters. For each pattern, equations describing pressure evolution were derived, along with two efficiency indices: hydraulic energy consumption per unit of filtered water and the ratio of filtered to total water used. Results show that reducing the clean filter pressure drop is the most effective way to improve energy efficiency: a 20% reduction decreases energy use per filtered volume by 7.6% under realistic conditions. Extending filtration time also improves energy efficiency for filters requiring high backwashing pressure, though less noticeably. Water use efficiency depends mainly on filtration duration relative to backwashing. Longer or higher‑flow backwashing substantially reduces efficiency; doubling backwashing time decreases water‑use efficiency by 4.5%. Optimal filter design should prioritize minimizing pressure loss to reduce energy demand and carefully controlling backwashing flow rate and duration to limit water losses.

Soil amendments can improve soil properties and crop yield, but their impact on agricultural product quality needs further investigation. A three-year field pot experiment was conducted using loamy sand soil to explore this issue. Green peppers were grown in the field under eight different soil treatments: Control (C); Biochar (B); Lime (L); Hydrogel (H); Biochar-Lime (BL); Biochar-Hydrogel (BH); Hydrogel-Lime (HL); and Biochar-Hydrogel-Lime (BHL). Biochar, lime (1% w/w), and hydrogel (0.5% w/w) were mixed with the soil to assess the quality of peppers at maturity. The amendments were arranged in a randomized complete block design (four blocks). Pepper fruits subjected to B, L, BH, HL, and BHL treatments showed the best diameter, length, thickness, volume, and dry matter content. Fruits treated with B, BH, and BHL exhibited higher firmness (6%-19%) and stiffness (16%-29%) than control and others. Treatments B, H, BH, and BHL significantly increased vitamin C content (20%-131%) in pepper fruits. The phenolic and flavonoid content increased significantly by 32%-139% and 27%-86%, respectively, when the soil was amended with B, H, BH, HL, and BHL than control. This study shows that soil amendments improve the quality of green peppers, offering nutritional value for consumers and economic benefits for producers.

Irrigated Agriculture in Morocco is facing a challenge of managing water demand and improving its efficiency and productivity. There is a need to establish cost-effective methods and workflows that can help farmers use evapotranspiration process based on AI-based irrigation management. This study aimed to build and evaluate a data driven process based on climatic data for modeling monthly reference evapotranspiration (ETo) at the arid land context of Chaouia-Morocco.

Daily meteorological data were treated to compute ETo using FAO-56 Penman–Monteith and ML methods. Daily ET0 data of 94 months were aggregated using a coverage threshold of 90% to show a good fitting of both outputs of daily (RMSE 0.131 mm d−1; R² 0.995) and monthly (RMSE 3.21 mm month−1; R² 0.996). The Random Forest method was trained on monthly data of mean temperature, relative humidity, wind speed, solar radiation, and precipitation to achieve a good fitting (RMSE 10.88 mm month−1; R² 0.962; r 0.989). The PCA (Principal component analysis) method showed that variance prediction can be explained by the main two components (86.8%) with a strong correlation of PC1 to monthly ET0 (r = 0.953). These results can be very useful for implementing a robust AI model for precise irrigation management.