Conference Agenda

Beyond Gray Infrastructure: Ecological Potential of Industrial areas for the Green Infrastructure Network

Sharon Esposito, Edoardo Asquini, Francesco Boscutti, Maurizia Sigura

Department of Agricultural, Food, Environmental and Animal Sciences, University of Udine

Green Infrastructure (GI) represents a fundamental tool for territorial sustainability, offering a strategic reinterpretation of industrial production areas. Although generally classified as "gray infrastructure" with a high environmental impact, these zones often possess a significant component of unbuilt green spaces which, if appropriately valued, can transform industrial sites into multifunctional nodes. This study proposes a multiscale methodological approach for mapping these surfaces using Remote Sensing and GIS technologies. At the local scale, the application of the Random Forest machine learning algorithm to Sentinel-2 satellite imagery enabled a supervised classification of land cover into herbaceous, shrubby, arboreal, and impervious surfaces. The model, trained via Regions of Interest was validated through direct field surveys. The integration of spectral variables with texture indices (GLCM) increased the classification accuracy enabling the effective discrimination between herbaceous surfaces and tree cover even within complex anthropized contexts. The results highlight environmental heterogeneity values in industrial areas, providing the quantitative basis for reconnecting production zones with nature thus making the concept of Green Infrastructure operational

Malaysia’s hot and humid climate challenges broiler production, leading to the adoption of low-cost double-skin roof (DSR) systems to improve thermal performance. The DSR consists of a secondary inclined roof above the primary roof, forming an air gap that reduces solar heat gain. However, computational fluid dynamics (CFD) simulations of DSR structure are computationally intensive due to complex heat transfer processes. This study proposes a hybrid modelling approach (HTM+CFD), in which the DSR is modelled using heat transfer models (HTM) and the indoor space is simulated using CFD. The HTM-estimated primary roof heat flux was applied as a boundary condition in the CFD model. The approach was evaluated under wet and dry season scenarios in a Malaysian commercial broiler house and validated using 24-hour measurements of air-gap temperatures and indoor temperature and air velocity at eighteen locations. The HTM+CFD were also compared with full-geometry DSR+CFD model and a CFD model without roof heat flux. The HTM+CFD and DSR+CFD models showed better agreement (CVRMSE < 3.0%, MAPE < 2.5%) than the model without roof heat flux. The HTM+CFD reduced computational cost by 50% while maintaining predictive accuracy, providing an efficient framework for thermal evaluation of DSR broiler houses in tropical climates.

Drainage wells are designed to lower groundwater levels and thereby prevent landslides. The 2024 Noto Peninsula Earthquake damaged the drainage wells, while field inspections of these structures are hazardous and time consuming, necessitating contactless assessment methods. Although InSAR derived from satellites such as Sentinel-1 has been widely applied to landslide monitoring, its potential for assessing the functional state of individual drainage wells has not been examined. A classification approach for distinguishing wells that experienced shear failure from undamaged wells is considered using Sentinel-1 time-series data acquired over the Noto Peninsula. Three drainage wells are examined: one confirmed to have experienced shear failure, and two confirmed to have retained drainage function following the earthquake. Two parameters are extracted from the Sentinel-1 time series: line-of-sight displacement, including decomposed vertical and horizontal components, and SAR backscatter-derived soil moisture as a proxy for subsurface water content. Time-series comparisons of these parameters between damaged and intact wells are conducted to evaluate whether systematic differences attributable to loss of drainage function are present.

. Livestock transportation is a critical component of Korea’s food supply chain, with approximately 3,000 cattle, 51,000 pigs, and 2.6 million chickens transported daily nationwide. During transport, animals are frequently exposed to heat stress, limited ventilation, poor air quality, vibration, and mechanical injury, resulting in substantial welfare and economic losses. Transport-related stress has been associated with increased mortality in chickens and with meat quality problems in cattle and pigs, including dark-coloured, firm beef and pale, soft pork with reduced water-holding capacity. In addition, emissions of odours, particulate matter, and pathogenic aerosols from transport vehicles pose risks to environmental safety, public health, and disease transmission. This study presents an integrated framework to assess welfare conditions and environmental risks during livestock transport by combining a vehicle microclimate prediction model based on physiological heat-exchange processes and computational fluid dynamics with video-based analytics for real-time detection of stress-related behaviours and injury indicators. Structural optimization and emergency cooling strategies for transport vehicles are evaluated alongside quantitative measurements of gaseous and particulate emissions. The framework is validated using sensor-equipped commercial vehicles operating under diverse meteorological and driving conditions, providing actionable insights to improve transport management, enhance animal welfare, and reduce avoidable losses.

Many terraced rice fields in Japan’s mountainous regions are irrigated by hillside irrigation canals, particularly in Kyushu. These canals are constructed along contour lines across mountain slopes and are often located within forested areas, resulting in more complex cross-sectional forms than canals in lowland plains. Although three-dimensional (3D) surveying using LiDAR-equipped smartphones has recently gained attention and has been applied to canal topographic measurement, no previous study has examined canals within forest environments. This study aimed to evaluate the feasibility of measuring hillside irrigation canals in forests using a handheld 3D surveying system. The system combined the LiDAR sensor of an iPhone 15 Pro Max with a viDoc RTK rover and point cloud data were processed using PIX4Dmatic. Surveys were conducted in two districts of Takachiho, Miyazaki Prefecture, Japan. Ground control points (GCPs) were installed, and longitudinal profiles were derived to assess positional differences between forward and return surveys. Results showed that GCPs reduced cumulative positional errors. Forest conditions decreased image acquisition and increased horizontal error; however, extending the measurement range improved image density and accuracy. Vegetation and wildlife fences hindered mesh generation. Despite variability in accuracy, the handheld system demonstrated potential for topographic measurement of forest hillside irrigation canals.

Heat stress is a major limiting factor affecting productivity and welfare of dairy cows in intensive production systems, particularly under tropical and subtropical climates. Effective mitigation requires continuous microclimate monitoring with high temporal and spatial resolution, integrated with data-driven management strategies. However, commercially available real-time environmental monitoring systems often involve high costs and limited scalability for on-farm implementation. In this context, a scalable IoT-based embedded architecture was developed, calibrated, and deployed for real-time thermal monitoring in a Compost-Bedded Pack Barn (CBP), targeting integration within Digital Livestock Farming systems. The platform comprised sensing modules equipped with BME280 sensors integrated with ESP32 microcontrollers and NRF24L01 wireless communication, supported by cloud-based data storage. Calibration in a climatic chamber showed high accuracy (R² > 0.98 for air temperature and relative humidity), followed by a ten-day deployment in a commercial CBP facility. The average transmission success rate was 95.94%, with 98.41% of data packets delivered within the one-minute target interval, demonstrating operational robustness under field conditions. Analyses indicated exposure to periods with air temperature > 24.00 °C and THI > 74.00. The proposed architecture shows strong potential for integration with automated environmental control systems, providing scalable decision support for precision livestock environmental management.

This study developed regression models to predict the internal temperature–humidity index in mechanically ventilated broiler houses using weather forecast data. A building energy simulation model was developed via TRNSYS and validated using field data, demonstrating high accuracy with mean absolute percentage error under 2.00%. To generate a robust dataset, 3,072 simulation cases were performed, considering various factors such as site, thermal transmittance, and cooling conditions. Based on ANOVA results that identified regional climate and cooling system operation as the most significant drivers of internal THI, three regression approaches were designed and compared: condition-specific equations, a unified model incorporating categorical predictors, and a single-variable model. The unified regression model provided the best performance (R² = 0.978, MAPE = 0.677%), offering a practical tool for real-time heat stress prediction. These results suggest that the developed model can be integrated with weather forecasts to provide early warnings and support decision-making for climate change adaptation in livestock production.

Many agricultural pipeline networks have exceeded its intended service lifespans, increasing the risk of water leakage and reduced conveyance efficiency. Detecting leaks over extensive systems remains challenging due to the buried and often inaccessible nature of the infrastructure. To address this, satellite remote sensing is investigated as a preliminary screening tool for leakage detection. The focus is placed on which satellite-derived spectral indices, including NDVI and other vegetation and moisture indicators, can capture surface changes associated with confirmed leakage events. Multi-temporal imagery was acquired over agricultural pipeline facilities where leakage had been recorded. Temporal difference images were computed for each index, and their spatial patterns were compared across pre- and post-leakage periods. Because the pipelines are far smaller than the pixel footprint of the sensors used, the pipeline is not imaged directly. Instead, the indirect surface response that leakage may produce in the surrounding area is traced. This study is grounded in field-confirmed leakage records from existing pipelines. Building upon these reference data, we aim to identify which indices most reliably distinguish leakage-affected conditions from normal background variability. This work positions satellite analysis as the opening stage of a tiered detection workflow, followed by field surveys and hydraulic diagnosis.

Agricultural water infrastructure requires continuous monitoring of both water availability and structural integrity. This study develops a satellite-based approach for agricultural dam water level estimation using NASA's ICESat-2 altimetry system and investigates correlations with dam displacement measurements. The methodology employs ICESat-2 ATLAS data processed through software to extract elevation measurements over dam reservoir surfaces. Satellite laser altimetry provides high-precision vertical measurements enabling detection of water level variations. Displacement monitoring is conducted separately to measure dam deformation. This study establishes correlations that could enable future displacement prediction from remotely sensed water level data alone by analyzing the relationship between satellite-derived water levels and measured dam displacements. Validation demonstrates multi-temporal monitoring capability with water level variations from 189 m to 203 m between 2019 and 2025. Analysis reveals correlations between water level fluctuations and dam displacement patterns. It suggests that potential for structural behavior prediction using only satellite observations. This relationship provides foundation for fully remote structural assessment. The satellite-based approach eliminates on-site instrumentation requirements while supporting both water resource management and dam safety evaluation. This cost-effective solution contributes to climate-resilient agricultural water supply through continuous remote monitoring.

Greenhouse production is highly energy-intensive and still largely dependent on fossil fuels. This work aims to define and apply a structured methodology for assessing and designing shallow geothermal systems coupled with ground-source heat pumps (GSHPs) to supply renewable heating energy to agricultural greenhouses

The study was carried out within the PRIN 2022 DiAGreen project on a typical three-bay greenhouse in Imola (Italy). Indoor temperature and relative humidity were digitally monitored and outdoor conditions were recorded by a weather station. An energy balance model was developed to estimate total heat demand, accounting for transmission losses (conduction and convection), ventilation losses (sensible and latent), radiative exchanges, and solar gains. Ground thermal properties were determined through a Thermal Response Test (TRT) on vertical probes.

Based on the available 460 m² installation area, two shallow geothermal solutions were evaluated: Basket Ground Heat Exchangers (BGHE) and Borehole Heat Exchangers (BHE), including a hybrid configuration integrated with the greenhouse climate control system. The analysis demonstrated that shallow geothermal systems can effectively cover greenhouse heating demand. The integration of GSHPs with environmental control systems represents a viable strategy for decarbonizing controlled environment agriculture and improving long-term energy sustainability.

Rural territories increasingly rely on low-impact infrastructures capable of fostering local development while preserving environmental and cultural resources. Long-distance cultural routes can be interpreted as linear rural infrastructures able to activate economic processes in marginal areas. This study assesses the socio-economic impacts generated by a section of the Via Francigena in Tuscany, focusing on its role as a territorial infrastructure connecting small settlements between Siena and southern Tuscany. The research integrates pedestrian flow data from a fixed monitoring station with spatial and statistical analysis of tourism accommodation performance. A GIS-based approach identified accommodation facilities within an 800 m buffer from the route, following a spatial criterion adopted by the Tuscany Region. Official data on tourist overnight stays were analysed to compare dynamics within and outside the buffer corridor. Flow data were combined with online survey responses to estimate expenditure patterns and average length of stay. Results show higher accommodation activity in settlements directly located along the route. The concentration of tourism presence within the 800 m corridor indicates a localized economic effect associated with pedestrian flows. Findings support the interpretation of the Via Francigena as a light rural infrastructure contributing to income diversification and service activation in marginal areas.

Reliability-based analysis is essential for structural design under uncertainty, particularly for agricultural greenhouse facilities that are highly vulnerable to variability in material properties, environmental loads, and connection behavior. However, practical reliability assessment is often limited by structural nonlinearity, unclear limit states, and the high computational cost of repeated finite element simulations. This study proposes an efficient reliability evaluation framework that combines an active learning-based deep neural network surrogate model with nonlinear finite element limit analysis. The finite element model captures complex structural behavior and defines safety-related limit states, while the deep learning meta-model significantly reduces computational demand during reliability estimation. Active learning is employed to enhance prediction accuracy with a minimal number of simulations. The methodology is applied to a standard greenhouse structure to evaluate its reliability index and safety performance. Results demonstrate that the proposed framework enables accurate and practical reliability assessment with substantially reduced computational effort, providing a rational basis for reliability-based safety design of agricultural greenhouse facilities.

This work was carried out with the support of “Research Program for Agriculture Science and Technology Development (Project No. RS-2025-02223042)” Rural Development Administration, Republic of Korea.

Smart livestock technologies increasingly address environmental sustainability, economic efficiency, and animal welfare challenges. However, research predominantly examines farmer adoption patterns, systematically overlooking technology manufacturers and providers whose development decisions and market perceptions critically shape adoption outcomes. This study bridges this research gap by investigating technological innovation strategies and perceived adoption barriers from the supply-side perspective. A structured survey was conducted with 68 manufacturers and providers across the dairy, swine, and poultry sectors. The survey examined integration of digital technologies, automation, IoT, and AI into product development, along with assessments of farm-level adoption dynamics, priority intervention areas, key drivers and barriers. Results reveal sector-specific priorities despite broad consensus on investment urgency within five years. Dairy technology providers identify animal welfare monitoring and precision feed management as critical needs; poultry manufacturers prioritise environmental control systems and health monitoring technologies; while pig sector stakeholders highlight the relevance of animal housing and facility infrastructure improvements, reflecting fundamental differences in production systems. By systematically documenting supply-side perspectives, this study reveals the alignment and gaps between technological innovation pathways and on-farm adoption realities. Findings provide actionable insights for targeting investments, designing adoption-support mechanisms, and formulating policies to accelerate the integration of smart technologies across livestock systems.

Digital monitoring of excavation progress during the construction of hillside farm ponds is crucial to ensure timely completion, cost reduction and improved operational safety in agricultural engineering projects. Traditional surveying methods, based on manual measurements and ground-based GNSS acquisition, require substantial manpower and expose operators to potential risks within active construction sites. The use of unmanned aerial systems (UAS) equipped with RGB sensors provides a promising alternative for high-resolution photogrammetric surveys. The scientific literature highlights several advantages of UAS-based photogrammetry and remote sensing, including very high spatial resolution, rapid data acquisition, mission flexibility and temporal repeatability, enabling multi-temporal analysis and change detection of earthworks. In this study, UAV-based surveys were conducted to generate orthomosaic and digital surface models for the estimation of net excavation volumes. The results were validated through ground-based GNSS measurements and cross-checked against soil transport data derived from truckload counts. UAV-derived volume calculations showed strong agreement with GNSS results, with deviations below 5%, and were consistent with logistical transport estimates. Moreover, survey time was reduced from several days to a few hours, significantly lowering labour costs and enhancing overall site safety. The findings confirm the technical reliability and operational efficiency of UAV-based photogrammetry for monitoring the construction of infrastructure.

The transition toward renewable energy and the reduction of fossil fuel dependence have become urgent priorities, driven by recent geopolitical dynamics and a heightened focus on environmental sustainability goals. In this context, agri-photovoltaic (APV) systems offer a synergistic solution to mitigate land-use conflicts. This study explores the integration of APV within marginal lands across Italy, identifying areas where physical marginality overlaps with high solar suitability. Such integration represents a strategic opportunity for socio-economic development and the valorization of territories often perceived only through their limitations, but remarkably resource-rich. The methodology employs deterministic and probabilistic approaches. First, a GIS analysis excluded restricted areas and high-value agricultural land. Second, a Spatial-MultiCriteria Decision Analysis (S-MCDA) was implemented to weigh factors such as erosion risk and solar exposure using an expert-based Analytic Hierarchy Process (AHP). The results of the deterministic approach highlight a high availability of land for APV, totaling 119,789 km². The probabilistic approach further refines this estimate, identifying high-suitability areas that involve 5% of the national surface (15,436 km²). These findings suggest a significant potential for strategic energy planning, providing a baseline for detailed local-scale assessments and crop-simulation models to optimize the co-production of energy and food on marginal lands.