Agriculture is among the main causes of water pollution. Currently, 75% of global anthropogenic nitrogen (N) loads come from leaching/runoff from cropland. The grey water footprint (GWF) is an indicator of water resource pollution, which allows for the evaluation and monitoring of pollutant loads (L) that can affect water. However, in the literature, there are different approaches to estimating L and thus contrasting GWF estimates: (A1) leaching/runoff fraction approach, (A2) surplus approach and (A3) soil nitrogen balance approach. This study compares these approaches for the first time to assess which one is best adapted to real crop production conditions and optimises GWF calculation. The three approaches are applied to assess N-related GWF in barley and soybean. For barley in 2019, A3 estimated a GWF value 285 to 196% higher than A1, while in 2020, the A3 estimate was 135 to 81% higher. Soybean did not produce a GWF due to the crop characteristics. A3 incorporated N partitioning within the agroecosystem and considered different N inputs beyond fertilization, improving the accuracy of L and GWF estimation. Providing robust GWF results to decision-makers may help to prevent or reduce the impacts of activities that threaten the world’s water ecosystems and supply.

Grasslands are extensively distributed in flatland areas around the world, such as the Pampas in South America. It is one of the most economically productive landscapes and, as in other regions, they are being replaced by forests at increasing rates. Soil salinization emerges as a negative consequence associated with water deficit and forest groundwater consumption (∼250–500 mm/yr, in this region). An assessment of forest groundwater consumption is crucial for risk evaluation of soil salinization on flatland environments. For this aim, numerical modeling based on physical/biological processes and atmospheric boundary conditions was successfully applied in monitored grassland and afforested plots. Modeling results suggested a partial hydraulic disconnection between forest and phreatic aquifer due to the presence of petrocalcic horizons. Forest transpiration estimates were approximately 13% of total groundwater usage. Forest water consumption was then restricted to that soil portions above the petrocalcic horizons. Estimated forest transpiration rates (∼723 mm/yr) were similar to and even exceeded those reported in salinized sites with similar features. However, the risk of salinization of these soils was unlikely, because forest transpiration was restricted to the upper soil portions filled with fresh rainwater. The petrocalcic horizon retained water and prevented both deep drainage and, indirectly, soil salinization.

Flatland areas, such as the Pampas in South America, rank among the world’s most economically productive landscapes. Over the last century, these regions have been increasingly used for intensive production, which has resulted in significant environmental impacts. These include alterations in pH, salinity, and other soil properties; changes in water flows and balances; increased floods and droughts; and heightened wind and water erosion. To address these challenges, numerical process-based models are essential to assess the highly variable, interconnected, and nonlinear processes that drive these impacts. Such models rely on comprehensive soil databases including hydraulic properties to provide representative results. This study aimed to develop a robust database of soil properties for the Buenos Aires Province in Argentina, encompassing much of the Pampas region. Using granulometric and physicochemical data from the Instituto Nacional de Tecnología Agropecuaria (INTA) database, we applied 38 pedotransfer functions to 381 soil profiles to estimate hydraulic parameters. These were compared with seven calibrated parameter sets from the different study sites. This study demonstrated that model performance varies depending on the evaluated function, with specific models excelling in particular variables, highlighting the need for careful selection based on the characteristics of the dataset. Las llanuras, como la Pampa en Sudamérica, se encuentran entre los paisajes más productivos del mundo. En el último siglo, estas regiones han sido intensamente explotadas para actividades agrícolas, generando impactos ambientales significativos, como alteraciones del pH y la salinidad del suelo, cambios en los flujos hídricos, aumento de inundaciones y sequías, y mayor erosión eólica e hídrica. Para abordar estos desafíos, los modelos numéricos basados en procesos son fundamentales para evaluar las interacciones complejas y no lineales que los impulsan. Estos modelos requieren bases de datos detalladas de propiedades del suelo, incluidas las hidráulicas. Este estudio desarrolló una base de datos de propiedades del suelo para la provincia de Buenos Aires (Argentina), que comprende gran parte de la región pampeana. A partir de datos granulométricos y fisicoquímicos del Instituto Nacional de Tecnología Agropecuaria (INTA), se aplicaron 38 funciones de pedotransferencia a 381 perfiles de suelo para estimar parámetros hidráulicos, comparándolos con siete conjuntos calibrados en siete sitios. Los resultados evidenciaron que el desempeño de las funciones varía según la variable evaluada, destacando la importancia de seleccionar modelos específicos según las características del conjunto de datos.