Salinization of soils and freshwater resources by natural processes and/or human activities has become an increasing issue that affects environmental services and socioeconomic relations. In addition, salinization jeopardizes agroecosystems, inducing salt stress in most cultivated plants (nutrient deficiency, pH and oxidative stress, biomass reduction), and directly affects the quality and quantity of food production. Depending on the type of salt/stress (alkaline or pH-neutral), specific approaches and solutions should be applied to ameliorate the situation on-site. Various agro-hydrotechnical (soil and water conservation, reduced tillage, mulching, rainwater harvesting, irrigation and drainage, control of seawater intrusion), biological (agroforestry, multi-cropping, cultivation of salt-resistant species, bacterial inoculation, promotion of mycorrhiza, grafting with salt-resistant rootstocks), chemical (application of organic and mineral amendments, phytohormones), bio-ecological (breeding, desalination, application of nano-based products, seed biopriming), and/or institutional solutions (salinity monitoring, integrated national and regional strategies) are very effective against salinity/salt stress and numerous other constraints. Advances in computer science (artificial intelligence, machine learning) provide rapid predictions of salinization processes from the field to the global scale, under numerous scenarios, including climate change. Thus, these results represent a comprehensive outcome and tool for a multidisciplinary approach to protect and control salinization, minimizing damages caused by salt stress.

Soil reflectance spectroscopy provides a rapid, non-destructive means of estimating soil properties, but the spectral resolution actually required for reliable prediction, and where within the spectrum that information resides, remains poorly quantified. We addressed this using 910, 731, and 416 Mediterranean soil samples characterised for clay, soil organic carbon (SOC), and calcium carbonate (CaCO₃), respectively. Laboratory hyperspectral reflectance data (400–2500 nm) was resampled into 5–100-band configurations and evaluated under both laboratory and simulated remote-sensing conditions, across five preprocessing treatments and five regression algorithms.Predictive performance plateaued well below full hyperspectral resolution: at 25 bands for clay (R² = 0.85–0.86 of an optimum of 0.89–0.90), 20 bands for SOC (R² = 0.44–0.48 of an optimum of 0.52–0.56), and 30 bands for CaCO₃ (R² = 0.51–0.53 of an optimum of 0.53–0.54), consistently across both scenarios. A second phase then seeded an exhaustive-then-greedy search with permutation-importance rankings to identify smaller, strategically located band subsets that matched this performance with only 5 bands for clay and 8–12 bands for SOC and CaCO₃ — over 99% fewer bands than the original hyperspectral resolution. These subsets substantially outperformed uniformly spaced bands of the same number (e.g. CaCO₃: R² = 0.53 vs. 0.05), indicating that band placement, not merely band count, determines predictive capacity.These results suggest that accurate prediction of clay, SOC, and CaCO₃ in Mediterranean soils does not necessarily require hyperspectral sampling, and offer quantitative targets for the number and placement of bands in future multispectral soil- and satellite-based sensing systems.

Process-based soil organic carbon (SOC) models are widely used for carbon accounting and assessments. For applications beyond experimental sites, remote sensing–derived inputs are often used to complement or replace in situ boundary conditions. This substitution introduces additional sensitivity that has rarely been explicitly quantified. Here, we propose a general framework to quantify boundary-condition sensitivity in SOC modelling, distinguishing between dynamic sensitivity, which reflects deviations in the temporal trajectory of SOC stocks, and endpoint sensitivity, which captures differences in estimated SOC changes over defined reporting periods. Sensitivity is quantified by comparing substitution simulations, in which boundary conditions are replaced one by one with remote sensing–derived spatial datasets, against a baseline simulation driven by in situ observations. The framework is applied to 16 long-term experimental sites across Europe, covering diverse land-use types and environmental conditions. Boundary-condition substitution affected both SOC trajectories and endpoint estimates, although the magnitude and form of these effects varied among inputs. Initial SOC, vegetation-derived carbon inputs, and climate data were the dominant drivers of dynamic sensitivity. Fixed endpoint sensitivity was generally small relative to measured SOC changes, supporting the use of remote sensing–derived spatial datasets as scalable boundary-condition proxies for SOC modelling. Rolling endpoint sensitivity further revealed that SOC change assessments depend not only on SOC dynamics but also on the selected reporting window. Short reporting periods were more affected by short-term variability, whereas longer periods were more susceptible to accumulated deviations. A reporting window of approximately 15–30 years provided a practical balance between these effects. Overall, the framework enables a systematic quantification of sensitivity arising from boundary-condition substitution, enabling evaluation of both trajectory and reporting robustness. These findings highlight the importance of jointly considering critical boundary conditions and reporting-window design when applying remote sensing–driven SOC modelling for regional carbon accounting.

Process-based soil organic carbon (SOC) models are widely used for carbon accounting and assessments. For applications beyond experimental sites, remote sensing–derived inputs are often used to complement or replace in situ boundary conditions. This substitution introduces additional sensitivity that has rarely been explicitly quantified. Here, we propose a general framework to quantify boundary-condition sensitivity in SOC modelling, distinguishing between dynamic sensitivity, which reflects deviations in the temporal trajectory of SOC stocks, and endpoint sensitivity, which captures differences in estimated SOC changes over defined reporting periods. Sensitivity is quantified by comparing substitution simulations, in which boundary conditions are replaced one by one with remote sensing–derived spatial datasets, against a baseline simulation driven by in situ observations. The framework is applied to 16 long-term experimental sites across Europe, covering diverse land-use types and environmental conditions. Boundary-condition substitution affected both SOC trajectories and endpoint estimates, although the magnitude and form of these effects varied among inputs. Initial SOC, vegetation-derived carbon inputs, and climate data were the dominant drivers of dynamic sensitivity. Fixed endpoint sensitivity was generally small relative to measured SOC changes, supporting the use of remote sensing–derived spatial datasets as scalable boundary-condition proxies for SOC modelling. Rolling endpoint sensitivity further revealed that SOC change assessments depend not only on SOC dynamics but also on the selected reporting window. Short reporting periods were more affected by short-term variability, whereas longer periods were more susceptible to accumulated deviations. A reporting window of approximately 15–30 years provided a practical balance between these effects. Overall, the framework enables a systematic quantification of sensitivity arising from boundary-condition substitution, enabling evaluation of both trajectory and reporting robustness. These findings highlight the importance of jointly considering critical boundary conditions and reporting-window design when applying remote sensing–driven SOC modelling for regional carbon accounting.

Soil reflectance spectroscopy is a powerful tool for rapid, non-destructive assessment of soil properties and the foundation for data-driven soil science applications. However, systematic discrepancies during routine spectral measurement procedures, particularly those arising from contamination or deterioration of white reference (WR) calibration panels, may compromise spectral data stability and hinder harmonization practices across laboratories. This study investigates the impact of using a non-contaminated WR panel as a calibration target to measure soil reflectance across the visible (Vis), near-infrared (NIR), and short-wavelength infrared (SWIR) spectral regions. The study evaluates the effectiveness of an internal soil standard (ISS) Lucky Bay sand to correct discrepancies within a controlled laboratory setting. Twelve soils from the Israeli Legacy Soil Spectral Library were analyzed using a contact-probe setup that was calibrated with both a clean and a contaminated WR. The spectral correction method, based on IEEE P4005 protocols and the ISS calibration, significantly reduced spectral inconsistencies, especially in the Vis region where contamination effects were most pronounced. Results show that the ISS effectively harmonized spectra acquired under different WR conditions, reducing the modified average spectral difference stability (mASDS) measure across all samples. While ISS correction is commonly employed for cross-laboratory harmonization, our findings highlight its critical role in enhancing intra-laboratory consistency under routine operational variability. We recommend that every WR calibration process will be accompanied with ISS measurements. The continuous use of a well-maintained WR and ISS improves the reliability of soil spectral datasets and supports the long-term harmonization of soil spectral libraries.