RESUMO
A spatiotemporal machine learning framework for automated prediction and analysis of long-term Land Use/Land Cover dynamics is presented. The framework includes: (1) harmonization and preprocessing of spatial and spatiotemporal input datasets (GLAD Landsat, NPP/VIIRS) including five million harmonized LUCAS and CORINE Land Cover-derived training samples, (2) model building based on spatial k-fold cross-validation and hyper-parameter optimization, (3) prediction of the most probable class, class probabilities and model variance of predicted probabilities per pixel, (4) LULC change analysis on time-series of produced maps. The spatiotemporal ensemble model consists of a random forest, gradient boosted tree classifier, and an artificial neural network, with a logistic regressor as meta-learner. The results show that the most important variables for mapping LULC in Europe are: seasonal aggregates of Landsat green and near-infrared bands, multiple Landsat-derived spectral indices, long-term surface water probability, and elevation. Spatial cross-validation of the model indicates consistent performance across multiple years with overall accuracy (a weighted F1-score) of 0.49, 0.63, and 0.83 when predicting 43 (level-3), 14 (level-2), and five classes (level-1). Additional experiments show that spatiotemporal models generalize better to unknown years, outperforming single-year models on known-year classification by 2.7% and unknown-year classification by 3.5%. Results of the accuracy assessment using 48,365 independent test samples shows 87% match with the validation points. Results of time-series analysis (time-series of LULC probabilities and NDVI images) suggest forest loss in large parts of Sweden, the Alps, and Scotland. Positive and negative trends in NDVI in general match the land degradation and land restoration classes, with "urbanization" showing the most negative NDVI trend. An advantage of using spatiotemporal ML is that the fitted model can be used to predict LULC in years that were not included in its training dataset, allowing generalization to past and future periods, e.g. to predict LULC for years prior to 2000 and beyond 2020. The generated LULC time-series data stack (ODSE-LULC), including the training points, is publicly available via the ODSE Viewer. Functions used to prepare data and run modeling are available via the eumap library for Python.
Assuntos
Monitoramento Ambiental , Urbanização , Probabilidade , Europa (Continente) , Fatores de TempoRESUMO
Although large amounts of pesticides are used annually and a majority enters the soil to form short- or long-term residues, extensive soil surveys for currently used pesticides (CUPs) are scarce. To determine the status of CUPs' occurrence in arable land in Central Europe, 51 CUPs and 9 transformation products (TPs) were analysed in 75 arable soils in the Czech Republic (CR) several months after the last pesticide application. Moreover, two banned triazines (simazine and atrazine) and their TPs were analysed because of their frequent detection in CR waters. Multi-residue pesticide analysis on LC-MS/MS after soil QuEChERS extraction was used. The soils contained multiple pesticide residues frequently (e.g. 51% soils with ≥5 pesticides). The levels were also noticeable (e.g. 36% soils with ≥3 pesticides exceeding the threshold of 0.01mg/kg). After triazine herbicides (89% soils), conazole fungicides showed the second most frequent occurrence (73% soils) and also high levels (53% soils with total conazoles above 0.01mg/kg). Frequent occurrence was found also for chloroacetanilide TPs (25% of soils), fenpropidin (20%) and diflufenican (17%). With the exception of triazines' negative correlation to soil pH, no clear relationships were found between pesticide occurrence and soil properties. Association of simazine TPs with terbuthylazine and its target crops proved the frequent residues of this banned compound originate from terbuthylazine impurities. In contrast, frequent atrazine-2-hydroxy residue is probably a legacy of high atrazine usage in the past. The occurrence and levels of compounds were closely associated with their solubility, hydrophobicity and half-life. The results showed links to CR water-monitoring findings. This study represents the first extensive survey of multiple pesticide residues in Central European arable soils, including an insight into their relationships to site and pesticide properties.