Spatial Coupling of Greenhouse Gas Emissions, Vegetation Activity, Groundwater Nitrate Occurrence, and Soil Environments Across Denmark: A 5-km Spatial Association and Soil-Stratified Analysis

Avideh Asadollahi 

Integrated environmental assessment increasingly requires the joint analysis of atmospheric emissions, vegetation dynamics, groundwater quality, and soil context rather than isolated thematic mapping. We developed a Denmark-wide 5-km analytical grid in EPSG:25832 containing 1,803 cells and harmonized four continuous environmental indicators: CO2 and CH4 emissions from EDGAR, vegetation activity represented by 2024 MODIS NDVI, and groundwater nitrate occurrence derived from national GEUS observations. Soil information from the GEUS Digital Soil Map was used to stratify environmental associations into five broad environmental groups. Global Pearson, Spearman, and Kendall associations were complemented by global Moran’s I, directional bivariate Moran’s I, local bivariate spatial association, an upper-1% CO2 sensitivity analysis, and soil-stratified statistics. CO2 and NDVI showed the strongest national association (Pearson r = −0.2621), while
CH4 was positively associated with CO2 (r = 0.0898) and nitrate occurrence (r = 0.0792). All four continuous variables exhibited significant positive spatial autocorrelation under both Queen and Rook contiguity, with the strongest structure for CH4 and CO2 and the weakest for nitrate. Bivariate Moran analysis supported a negative CO2–NDVI cross-association and positive CH4–CO2 and CH4–nitrate cross-associations. Soil-stratified results showed substantial heterogeneity; the strongest Pearson CO2–NDVI association occurred in Human-affected & Technological environments (r = −0.6208), whereas Sandy & Poor and Sedimentary & Clay-rich environments displayed the clearest positive CH4–CO2 associations. The study is explicitly associative rather than causal. Its principal contribution is a spatially explicit national screening framework that reveals where environmental indicators co-occur, where relationships are spatially organized, and where targeted process-based investigation and monitoring may be most informative.