Search

In this work, we tested different combinations of sampling strategies, random sampling and conditioned Latin Hypercube sampling (cLHS)] and sample ratios (10% = 147 and 25% = 368) to predict soil phosphorus and potassium contents, previously estimated using standard laboratory protocols. Other environmental covariates, used as input data for prediction, were obtained from different sources (multispectral Landsat-OLI 8 image, WorldClim database, ISRIC soil database, and ASTER-GDEM). Our findings showed that random sampling was suitable for predicting phosphorus, while the conditioned Latin Hypercube sampling was suitable for predicting potassium. Furthermore, we observed that when the sample ratio increased from 10 to 25%, model accuracy improved in random sampling and cLHS for phosphorus and potassium prediction. However, before generalizing these findings, we recommend that further studies be conducted under different conditions (climate, soil types and parent materials) and testing other sample ratios to determine the best sampling strategy with the optimum ratio to predict soil nutrients better. [ABSTRACT FROM AUTHOR]

• Hybridized digital soil mapping model based on machine learning was developed. • The hybrid model was compared to Cokriging (Cok) and Gaussian process regression (GPR). • The hybrid method has, in general, better performance than the Cok and GPR. As a widely used soil mapping method, the kriging method involves a high sampling point to generate quality and accurate maps. Combining kriging and machine learning (ML) can produce soil maps with fewer number sampling points. This study's objective was to implement a hybrid approach based on the Cokriging (Cok) and an ML technique [i.e., Gaussian process regression (GPR)]. The hybrid method (called the Cok-GPR method) uses the Cok (Coki, i = 1 to n) as a predictor method of the soil sulphur and then uses GPR to improve the prediction accuracy. The proposed method was compared with the Cok and the GPR models, respectively, in a case study. Soil samples (n = 115) were collected from the topsoil (0–20) at the agricultural site of approximately 889.8 km2 size. S, Ca, K, Mg, Na, P, and V were estimated via Inductively Coupled Plasma Optical Emission Spectroscopy (ICP-OES) equipment and presented as S_ICP-OES (response variable), and predictors (Ca_ICP-OES, K_ICP-OES, Mg_ICP-OES, Na_ICP-OES, P_ICP-OES, and V_ICP-OES), respectively. For GPR and Cok-GPR, an 80% (calibration) to 20% (validation) random dataset split was performed. The calibration dataset was implemented under k = 10-fold cross-validation, repeated five times. All the models were evaluated by MAE, RMSE and R2 criteria. According to the model and map performances. Cok1 model via Ca_ICP-OES, K_ICP-OES, Mg_ICP-OES gave the best model (MAE = −1.28 mg/kg RMSE = 164.42 mg/kg, R2 = 0.85). Its corresponding GPR1 approach, modelled with the same predictors produced the best (MAE = 85.43 mg/kg, RMSE = 137.59 mg/kg, R2 = 0.83). While the hybrid Cok1-GPR model produced MAE = 76.84 mg/kg, RMSE = 102.11 mg/kg, and R2 = 0.91. The model outperformed both the Cok and GPR models, respectively. The proposed Cok-GPR model can be applied to efficiently predict soil nutrient element levels at the regional level and be useful during policymaking. [ABSTRACT FROM AUTHOR]