Search

Storm ‘Daniel’ caused the most severe flood phenomenon that Greece has ever experienced, with thousands of hectares of farmland submerged for days. This led to sediment deposition in the inundated areas, which significantly altered the chemical properties of the soil, as revealed by extensive soil sampling and laboratory analysis. The causal relationships between the soil chemical properties and sediment deposition were extracted using the DirectLiNGAM algorithm. The results of the causality analysis showed that the sediment deposition affected the CaCO3 concentration in the soil. Also, causal relationships were identified between CaCO3 and the available phosphorus (P-Olsen), as well as those between the sediment deposit depth and available manganese. The quantified relationships between the soil variables were then used to generate data using a Multiple Linear Perceptron (MLP) regressor for various levels of deposit depth (0, 5, 10, 15, 20, 25, and 30 cm). Then, linear regression equations were fitted across the different levels of deposit depth to determine the effect of the deposit depth on CaCO3, P, and Mn. The results revealed quadratic equations for CaCO3, P, and Mn as follows: 0.001XCaCO32 + 0.08XCaCO3 + 6.42, 0.004XP2 − 0.26XP + 12.29, and 0.003XMn2 − 0.08XMn + 22.47, respectively. The statistical analysis indicated that corn growing in soils with a sediment over 10 cm requires a 31.8% increase in the P rate to prevent yield decline. Additional notifications regarding cropping strategies in the near future are also discussed.

Machine learning (ML) algorithms are extensively used with outstanding prediction accuracy. However, in some cases, their overfitting capabilities, along with inadvertent biases, might produce overly optimistic results. Spatial data are a special kind of data that could introduce biases to ML due to their intrinsic spatial autocorrelation. To address this issue, a special resampling method has emerged called spatial cross-validation (SCV). The purpose of this study was to evaluate the performance of SCV compared with conventional random cross-validation (CCV) used in most ML studies. Multiple ML models were created with CCV and SCV to predict groundwater electrical conductivity (EC) with data (A) from Rhodope, Greece, in the summer of 2020; (B) from the same area but at a different time (summer 2019); and (C) from a new area (the Salento peninsula, Italy). The results showed that the SCV provides ML models with superior generalization capabilities and, hence, better prediction results in new unknown data. The SCV seems to be able to capture the spatial patterns in the data while also reducing the over-optimism bias that is often associated with CCV methods. Based on the results, SCV could be applied with ML in studies that use spatial data.

One of the core tasks in digital soil mapping (DSM) studies is the estimation of the spatial distribution of different soil variables. In addition, however, assessing the uncertainty of these estimations is equally important, something that a lot of current DSM studies lack. Machine learning (ML) methods are increasingly used in this scientific field, the majority of which do not have intrinsic uncertainty estimation capabilities. A solution to this is the use of specific ML methods that provide advanced prediction capabilities, along with innate uncertainty estimation metrics, like Quantile Regression Forests (QRF). In the current paper, the prediction and the uncertainty capabilities of QRF, Random Forests (RF) and geostatistical methods were assessed. It was confirmed that QRF exhibited outstanding results at predicting soil organic matter (OM) in the study area. In particular, R2 was much higher than the geostatistical methods, signifying that more variation is explained by the specific model. Moreover, its uncertainty capabilities as presented in the uncertainty maps, shows that it can also provide a good estimation of the uncertainty with distinct representation of the local variation in specific parts of the area, something that is considered a significant advantage, especially for decision support purposes.

A classic method for assessing the reference crop evapotranspiration (ETo) is the pan evaporation (Epan) method that uses Epan measurements and pan coefficient (kp) models, which can be functions of relative humidity (RH), wind speed (u2), and temperature (T). The aim of this study is to present a methodology for evaluating the robustness of regression coefficients associated to climate parameters (RH, u2, and T) in pan method models during recalibration at local conditions. Two years of daily data from April to October (warm season) of meteorological parameters, Epan measurements from class A pan evaporimeter and ETo estimated by ASCE-standardized method for the climatic conditions of Thessaloniki (Greece, semi-arid environment), were used. The regression coefficients of six general nonlinear (NLR) regression Epan models were analyzed through recalibration using a technique called “random cross-validation nonlinear regression RCV-NLR” that produced 1000 random splits of the initial dataset into calibration and validation sets using a constant proportion (70% and 30%, respectively). The variance of the regression coefficients was analyzed based on the 95% interval of the highest posterior density distribution. NLR models that included coefficients with a 95% HPD interval that fluctuates in both positive and negative values were considered nonrobust. The machine-learning technique of random forests (RF) was also used to build a RF model that includes Epan, u2, RH, and T parameters. This model was used as a benchmark for evaluating the predictive accuracy of NLR models but, also, for assessing the relative importance of the predictor climate variables if they were all included in one NLR model. The findings of this study indicated that locally calibrated NLR functions that use only the Epan parameter presented better results, while the inclusion of additional climate parameters was redundant and led to underfitting.

In the current paper we assess different machine learning (ML) models and hybrid geostatistical methods in the prediction of soil pH using digital elevation model derivates (environmental covariates) and co-located soil parameters (soil covariates). The study was located in the area of Grevena, Greece, where 266 disturbed soil samples were collected from randomly selected locations and analyzed in the laboratory of the Soil and Water Resources Institute. The different models that were assessed were random forests (RF), random forests kriging (RFK), gradient boosting (GB), gradient boosting kriging (GBK), neural networks (NN), and neural networks kriging (NNK) and finally, multiple linear regression (MLR), ordinary kriging (OK), and regression kriging (RK) that although they are not ML models, they were used for comparison reasons. Both the GB and RF models presented the best results in the study, with NN a close second. The introduction of OK to the ML models’ residuals did not have a major impact. Classical geostatistical or hybrid geostatistical methods without ML (OK, MLR, and RK) exhibited worse prediction accuracy compared to the models that included ML. Furthermore, different implementations (methods and packages) of the same ML models were also assessed. Regarding RF and GB, the different implementations that were applied (ranger-ranger, randomForest-rf, xgboost-xgbTree, xgboost-xgbDART) led to similar results, whereas in NN, the differences between the implementations used (nnet-nnet and nnet-avNNet) were more distinct. Finally, ML models tuned through a random search optimization method were compared with the same ML models with their default values. The results showed that the predictions were improved by the optimization process only where the ML algorithms demanded a large number of hyperparameters that needed tuning and there was a significant difference between the default values and the optimized ones, like in the case of GB and NN, but not in RF. In general, the current study concluded that although RF and GB presented approximately the same prediction accuracy, RF had more consistent results, regardless of different packages, different hyperparameter selection methods, or even the inclusion of OK in the ML models’ residuals.

The aim of this study is to develop an integrated approach to soil quality and fertility assessment in high-yielding rice agro-ecosystems threatened due to overexploitation of soil resources by intensive agriculture. The proposed approach is implemented considering representative pilot fields allocated throughout a study area based on the assumption that soils of similar general properties present a similar nutritional status due to common long-term management practices. The analysis includes (a) object-based image analysis for land zonation, (b) hot-spot analysis for sampling scheme evaluation, (c) setting of critical thresholds in soil parameters for detecting nutrient deficiencies and soil quality problems, and (d) Redundancy Analysis, TITAN analysis, and multiple regression for identifying individual or combined effects of general soil properties (e.g., organic matter, soil texture, pH, salinity) or non-soil parameters (e.g., topographic parameters) on soil nutrients. The approach was applied using as a case study the large rice agro-ecosystem of Thessaloniki plain in Greece considering some site specificities (e.g., high rice yields, calcareous soils) when setting the critical thresholds in soil parameters. The results showed that (a) 62.5% of the pilot fields’ coverage has a simultaneous deficiency in Zn, Mn, and B, (b) organic matter (OM) was the most significant descriptor of nutrients’ variance, and its cold spots (clustered regions of low OM values) showed important overlapping with the cold spots of K, Mg, Zn, Mn, Cu, and B, (c) a higher rate of availability increase in P, K, Mg, Mn, Zn, Fe, Cu, and B was observed when the OM ranged between 2 and 3%, and (d) the multiple regression models that assess K and P concentrations based on general soil properties showed an adequate performance, allowing their use for general assessment of their soil concentrations in the fields of the whole agro-ecosystem.

In this study, different interpolation techniques are presented, assessed, and compared for the estimation of soil iron (Fe) contents in locations where observations were not available. Initially, 400 soil samples from the Kozani area, which is near Polifitou Lake in northern Greece, were randomly collected from 2013 to 2015 and were analysed in the laboratory to determine the soil Fe concentrations and pH. The soil Fe concentrations were examined for spatial autocorrelation, and semivariograms were used to determine whether pH and Fe exhibited spatial cross correlation. Three interpolation methods, including Ordinary Kriging, Universal Kriging, and Co-Kriging, were applied, and their results were compared with the use of two different cross-validation methods. In the current study, there was evidence of spatial cross correlation of soil Fe and pH for each year, which was subsequently used to improve the interpolation results in locations where there were no measurements. In nearly all cases, Co-Kriging, which takes advantage of the covariance between the two regionalized variables (Fe and pH), outperformed the other interpolation techniques each year.