Mapping soil organic matter and identifying potential controls in the farmland of Southern China: Integration of multi‐source data, machine learning and geostatistics

Abstract

Soil organic matter (SOM) plays a critical role in terrestrial ecosystem functioning and is closely related to many global issues like soil fertility, soil health and climate regulation. Therefore, obtaining accurate information on the spatial distribution of SOM and its potential controlling factors is of global interest. However, this remains a great challenge since SOM is affected by numerous natural and anthropogenic factors and usually showed strong heterogeneity. In this study, we collected a total of 16,580 surface soil (0–20 cm) samples from the farmland throughout Jiangxi Province. And the Random Forest (RF), Cubist and gradient‐boosted models were compared and used to define the factor which is most associated with SOM. Then the ordinary kriging (OK) and machine learning‐ordinary co‐kriging (ML‐COK) were used to map SOM. We found that on average, 30.86 g kg−1 SOM was present in farmland soil of Jiangxi Province. Anthropogenic activities strongly affected SOM level, with five of the top 10 most important factors are anthropogenic related. The straw return amount was proved to have the largest importance (31.46%) for modelling SOM and a significant (p < 0.001) positive relationship between SOM content and the amount of straw returned to farmland was detected. Additionally, returning straw improved crop production. Soil derived from the Quaternary Subred Sand has the highest SOM content (37.82 g kg−1). Crop rotation also improved SOM content and the rice‐bean rotation system has the highest SOM content (34.27 g kg−1). With the best performance, the RF algorithm (R2 = 0.49, RMSE = 6.77 g kg−1) was selected to identify the primary control of SOM and integrated with COK, which we termed as ML‐COK, to map the SOM in the farmland of Jiangxi Province. ML‐COK outperformed OK method for mapping the SOM in farmland of Jiangxi Province with R2 of 0.351 and Lin's concordance correlation coefficient of 0.549. Farmland distributed in the central part of the province had high SOM content. In contrast, farmland in the north, south and east parts had relatively low SOM. Our study offers new insight for mapping soil properties, identifying potential factors driving variation in SOM, and also provides valuable information for making more reasonable and environmentally friendly farmland management measures.

Mapping Soil Organic Matter Using Different Modeling Techniques in the Dryland Agroecosystem of Huang-Huai-Hai Plain, Eastern China

“…In this study, K-nearest neighbor analysis is used to determine the expansion constants from different centers, which overcomes the shortage of the same weight of environmental variables in different spatial directions in the traditional model, which describes the local information in detail and effectively improves the prediction, accuracy, and generalization ability [25]. Our study demonstrated that the RF model achieved the highest prediction accuracy for mapping the SOM of farmland in Huang-Huai-Hai Plain, which is consistent with the farmland in southwest China [7], northeast and North Plain China [24], and Nepal [20]. This is owing to the RF having fewer constraints on sample collection and the ability to deal with robust outliers [12,48,64].…”

“…Furthermore, SOM is tightly linked to human activities in dryland agroecosystems, and covariates characterizing human activities such as tillage systems and fertilization methods were not included in the model predictions due to limited data acquisition. The practice has proven that agricultural activities such as rotation, irrigation, and fertilization significantly impact SOM [3,7,65]. Therefore, finding more auxiliary variables with a strong correlation with the SOM and alternative factors that can represent human activities as model inputs will be among the important ways in which to improve the accuracy of SOM prediction.…”

“…and anthropogenic (fertilization, rotation system, etc.) factors, showing an uneven spatial distribution, which directly affects the soil quality and crop yield in an agroecosystem [7][8][9]. Therefore, obtaining high-precision spatial distribution of SOM in farmland is the key to improving agricultural productivity and is still a challenging task.…”

Aboveground Biomass Component Plasticity and Allocation Variations of Bamboo (Pleioblastus amarus) of Different Regions

“…The PLS-SEM was used to explore composite factors affecting biomass variation and associative interactions and to understand regional biomass variation in driving factors using Smart PLS 4.0.9.4 [57]. Topographical sampling maps were generated using ArcGIS 10.7 (ESRI ® ArcMap™ 10.7) [58].…”