Modeling arsenic in European topsoils with a coupled semiparametric (GAMLSS-RF) model for censored data.

Arsenic (As) is a versatile heavy metalloid trace element extensively used in industrial applications. As is carcinogen, poses health risks through both inhalation and ingestion, and is associated with an increased risk of liver, kidney, lung, and bladder tumors. In the agricultural context, the repeated application of arsenical products leads to elevated soil concentrations, which are also affected by environmental and management variables. Since exposure to As poses risks, effective assessment tools to support environmental and health policies are needed. However, the most comprehensive soil As data available, the Land Use/Cover Area frame statistical Survey (LUCAS) database, contains severe limitations due to high detection limits. Although within International Organization for Standardization standards, the detection limits preclude the adoption of standard methodologies for data analysis. The present work focused on developing a new method to model As contamination in European soils using LUCAS soil samples. We introduce the GAMLSS-RF model, a novel approach that couples Random Forests with Generalized Additive Models for Location, Scale, and Shape. The semiparametric model can capture non-linear interactions among input variables while accommodating censored and non-censored observations and can be calibrated to include information from other campaign databases. After fitting and validating a spatial model, we produced European-scale As concentration maps at a 250 m spatial resolution and evaluated the patterns against reference values (i.e., two action levels and a background concentration). We found a significant variability of As concentration across the continent, with lower concentrations in Northern countries and higher concentrations in Portugal, Spain, Austria, France and Belgium. By overcoming limitations in existing databases and methodologies, the present approach provides an alternative way to handle highly censored data. The model also consists of a valuable probabilistic tool for assessing As contamination risks in soils, contributing to informed policy-making for environmental and health protection.

Spatial assessment of topsoil zinc concentrations in Europe.

Zinc (Zn) is essential to sustain crop production and human health, while it can be toxic when present in excess. In this manuscript, we applied a machine learning model on 21,682 soil samples from the Land Use and Coverage Area frame Survey (LUCAS) topsoil database of 2009/2012 to assess the spatial distribution in Europe of topsoil Zn concentrations measured by aqua regia extraction, and to identify the influence of natural drivers and anthropogenic sources on topsoil Zn concentrations. As a result, a map was produced showing topsoil Zn concentrations in Europe at a resolution of 250 m. The mean predicted Zn concentration in Europe was 41 mg kg-1, with a root mean squared error of around 40 mg kg-1 calculated for independent soil samples. We identified clay content as the most important factor explaining the overall distribution of soil Zn in Europe, with lower Zn concentrations in coarser soils. Next to texture, low Zn concentrations were found in soils with low pH (e.g. Podzols), as well as in soils with pH above 8 (i.e., Calcisols). The presence of deposits and mining activities mainly explained the occurrence of relatively high Zn concentrations above 167 mg kg-1 (the one percentile highest concentrations) within 10 km from these sites. In addition, the relatively higher Zn levels found in grasslands in regions with high livestock density may point to manure as a significant source of Zn in these soils. The map developed in this study can be used as a reference to assess the eco-toxicological risks associated with soil Zn concentrations in Europe and areas with Zn deficiency. In addition, it can provide a baseline for future policies in the context of pollution, soil health, human health, and crop nutrition.

From regional to parcel scale: A high-resolution map of cover crops across Europe combining satellite data with statistical surveys.

The reformed Common Agricultural Policy of 2023-2027 aims to promote a more sustainable and fair agricultural system in the European Union. Among the proposed measures, the incentivized adoption of cover crops to cover the soil during winter provides numerous benefits such as improved soil structure and reduced nutrient leaching and erosion. Despite this recognized importance, the availability of spatial data on cover crops is scarce. The increasing availability of field parcel declarations in the European Union has not yet filled this data gap due to its insufficient information content, limited public availability and a lack of standardization at continental scale. At present, the best information available is regionally aggregated survey data, which although indicative, hinders the development of spatially accurate studies. In this work, we propose a statistical model relating Sentinel-1 data to the existence of cover crops at the 100-m spatial resolution over the entirety of the European Union and United Kingdom and estimate its parameters using the spatially aggregated survey data. To validate the method in a spatially-explicit way, predictions were compared against farmers' registered declarations in France, where the adoption of cover crops is widespread. The results indicate a good agreement between predictions and parcel-level data. When interpreted as a binary classifier, the model yielded an Area Under the Curve (AUC) of 0.74 for the whole country. When the country was divided into five regions for the evaluation of regional biases, the AUC values were 0.77, 0.75, 0.74, 0.70, and 0.65 for the North, Center, West, East, and South regions respectively. Despite limitations such as the lack of data for validation outside France, and the non-standardized nomenclature for cover crops among Member States, this work constitutes the first effort to obtain a relevant cover crop map at a European scale for researchers and practitioners.

Disclosing contrasting scenarios for future land cover in Brazil: Results from a high-resolution spatiotemporal model.

Gaining information on the dynamics of land cover changes is a valuable step towards improving practical conservation actions. In recent years, the Brazilian presidential elections in 2018 and the recovery from one of the nation's worst economic recessions defined a political scenario that has been causing shifts in the patterns of land cover change. A variety of national plans for the near-future exist and include the construction of new roads connecting remote Amazonian areas and large dams that could flood up to 10 million hectares. These development plans threaten environmental conservation, but the potential effects on the local or regional land cover are mostly unknown. In this work, we construct a model to evaluate the possible consequences of policy actions on land cover dynamics in the near-future at a high-resolution scale. The regression model extracts the historical relationships between land cover and spatial drivers of change, and its extrapolation for the future enables the simulation of scenarios for the national plans currently discussed in Brazil. We also simulate three scenarios based on the Representative Concentration Pathways of the Intergovernmental Panel on Climate Change, which makes contrasting management assumptions. The resulting maps indicate that considerable changes in land cover composition and configuration may occur even in a short period. The historical Brazilian economic forces make the decrease in natural vegetation probabilities challenging to stop even in an environmentally oriented scenario, where plans for the construction of new infrastructure are abruptly interrupted. Our results also indicate that environmental degradation cannot be prevented without coordinated efforts between public agencies with a broad diversity of development viewpoints.