Effects of land use types and environmental factors on spatial distribution of soil total nitrogen in a coalfield on the Loess Plateau, China

Soil total nitrogen (TN) is one of the most important nutrients for plant and crop growth. It is essential to estimate the spatial distribution of TN to evaluate soil productivity, land reclamation efficiency, and agricultural management. This study was performed to investigate the spatial distribution of TN and its influencing factors in a coalfield in the Loess Plateau, China. A total of 143 topsoil samples were collected from five land use types and different slope positions. Classical statistical and geostatistical methods were used to quantify the spatial variation in TN. Redundancy analysis (RDA) was used to detect correlations between environmental factors and TN, whereas the random forest (RF) algorithm and LightGBM were applied to simulate the correlations. The results showed that the degree of spatial dependence ranged from 31.7%-45.2%, indicating moderate spatial variability in TN. Land use type significantly affected TN. The highest TN contents were found in the conservation tillage lands, followed by conventional tillage land, shrub land, grassland, and coal mining areas. The spatial distribution of the TN was also influenced by the slope position. The sites located downslope had higher TN than those located on the upper and middle slopes. RDA indicated that available potassium (AK) and soil organic matter (SOM) significantly affected the spatial distribution of TN. It was also demonstrated by LightGBM and RF that the TN content could be simulated by AK, SOM, normalised vegetation index, elevation, and available phosphorus. These results are helpful in understanding the effects of anthropogenic and natural factors on soil TN in coalfield areas.