Soil organic carbon (SOC) is a key component of the terrestrial carbon cycle and is essential for maintaining soil quality and fertility. Current European Union regulations call for accurate SOC spatial modelling based on sound sampling protocols. Reliable estimations of soil organic carbon stocks (SOCS) critically depend on an appropriate field sampling design, yet the influence of sampling protocols on SOC quantification remains poorly scrutinized. This study evaluates how two widely used sampling approaches, the Spanish National protocol and the harmonised LUCAS Soil protocol, affect SOC concentration, spatial variability and SOC stock estimates across the heterogeneous landscapes of the Autonomous Region of Madrid (Spain). Soil samples were collected at identical locations using both protocols, enabling a direct comparison of SOC measurements and their geostatistical representation. Mean SOC concentrations differed by less than 5% between protocols, but SOCS estimates diverged markedly (67 vs. 83.4 Mg C ha⁻¹), resulting in a regional difference of approximately 16 Tg of carbon. Semivariogram analyses revealed distinct spatial structures, with the National protocol capturing broader spatial dependence and greater local heterogeneity, while LUCAS produced narrower ranges and slightly lower minimum values. Despite these differences, both datasets reproduced consistent regional patterns, with low SOC in cultivated areas and higher SOC in forested sectors of the Guadarrama mountainous range. The results demonstrate that sampling design is a major source of uncertainty in SOCS estimation and highlight the need for harmonised sampling strategies to support carbon accounting in climate policy frameworks.
18 March 2026 in Congresos UEX, actas de congresos
El empleo de imágenes y sensores remotos pueden ser un apoyo importante para el fenotipado de cultivos. El objetivo de este estudio fue explorar la capacidad de medidas proximales y remotas para estimar la estimación de rendimiento y las diferencias en fenología de 57 genotipos de trigo procedentes de diversos países europeos. Para ello se seleccionaron 57 variedades de un ensayo que se llevó a cabo durante dos campañas de cultivo en Alcalá de Henares (Madrid). De cada variedad se determinaron las fechas de espigado, floración y maduración, así como el rendimiento en cosecha. Igualmente se tomaron medidas con un sensor de proximidad para calcular los índices NDVI, NDRE y CCCI así como imágenes de campo para obtener la cubierta vegetal en tres fechas. La altura del cultivo se obtuvo a partir de imágenes aéreas adquiridas con un UAV. La variable que mostró mayor capacidad para estimar el rendimiento fue la cubierta vegetal que, al incluir la altura estimada del dosel, permitió alcanzar un R² de 0,53 y un RMSE de 990 kg/ha. Al analizar las variedades según su origen, se observó que las mejoras más significativas correspondieron a las variedades españolas y suizas, con valores de R² de 0,43 y 0,53, respectivamente. En cuanto a la fecha de adquisición, la variable que mejor estimó el momento de espigado fue el índice CCCI, con un error del 47 %. Asimismo, los índices espectrales y los rasgos estructurales del cultivo aportaron información complementaria para explicar el rendimiento. Como alternativa basada en teledetección, se propone el uso de modelos texturales o geométricos del cultivo y técnicas de inteligencia artificial, que han demostrado ser eficaces en la estimación de la maduración de frutos y en la identificación de malas hierbas.
Soil organic carbon (SOC) is a key component of the terrestrial carbon cycle and is essential for maintaining soil quality and fertility. Current European Union regulations call for accurate SOC spatial modelling based on sound sampling protocols. Reliable estimations of soil organic carbon stocks (SOCS) critically depend on an appropriate field sampling design, yet the influence of sampling protocols on SOC quantification remains poorly scrutinized. This study evaluates how two widely used sampling approaches, the Spanish National protocol and the harmonised LUCAS Soil protocol, affect SOC concentration, spatial variability and SOC stock estimates across the heterogeneous landscapes of the Autonomous Region of Madrid (Spain). Soil samples were collected at identical locations using both protocols, enabling a direct comparison of SOC measurements and their geostatistical representation. Mean SOC concentrations differed by less than 5% between protocols, but SOCS estimates diverged markedly (67 vs. 83.4 Mg C ha⁻¹), resulting in a regional difference of approximately 16 Tg of carbon. Semivariogram analyses revealed distinct spatial structures, with the National protocol capturing broader spatial dependence and greater local heterogeneity, while LUCAS produced narrower ranges and slightly lower minimum values. Despite these differences, both datasets reproduced consistent regional patterns, with low SOC in cultivated areas and higher SOC in forested sectors of the Guadarrama mountainous range. The results demonstrate that sampling design is a major source of uncertainty in SOCS estimation and highlight the need for harmonised sampling strategies to support carbon accounting in climate policy frameworks.