Incineration of municipal sludge and agri-food by-products generates large quantities of ash that can be used in agriculture as phosphorus fertilizer. The fertilizing potential of sludge incineration ash (SIA) from 12 cities in Canada and the United States was tested in a greenhouse experiment against a synthetic fertilizer (TSP: triple superphosphate), a natural fertilizer (RP: rock phosphate), and a control without any P fertilizer. Two soil types were used: clay and sandy loam. A reliable a priori indicator of SIA P bioavailability was determined using the random forest method. SIA application increased ryegrass P uptake. The SIA relative P effectiveness (RPE), compared to the TSP, varied from 5.1% to 46.2% depending on the sludge origin and P solubility. SIA RPE was greater than RP for the clay soil but similar for the sandy loam soil. The neutral ammonium citrate (NAC) extraction, sometimes inappropriately used to characterize P availability of sludge and by-products, explained only 53% of the RPE variation. The random forest analysis showed that the oxalate extraction (Al, P, and Fe) is a better indicator (R² = 0.94) of relative availability of SIA than the NAC P solubility (R² = 0.86), and that Al content is the factor that influences most SIA P solubility. Based on our findings, we recommend the use of the Al, Fe, and P oxalate extraction to predict the SIA P availability, instead of the widely used NAC method which extracts only P.

Several studies have pointed out the promising use of nutritional diagnosis methods for the determination of optimum nutrient contents in plant tissues. The present investigation was carried out in different oases in Southern Tunisia to determine reference values for the interpretation of leaf analyses of date palm (Phoenix dactylifera) Deglet Nour cultivar with the Critical Value Approach (CVA) and the Compositional Nutrient Diagnosis (CND). A database (n = 100) of yield and mineral concentrations taken from date palm leaflets in October, at the maturity stage of dates, was used. The yield cut-off between low-yield and high-yield subpopulations, selected from cumulative variance ratio functions across survey data, was 76 kg palm⁻¹ and the global nutrient imbalance index (CNDr²) was 10.06. Critical CND nutrient indices were found to be symmetrical around zero as follows: (1.59; +1.59) for I_N, (−0.44, +0.44) for I_P, (−0.63, +0.63) for I_K, (−0.94, +0.94) for I_Ca, (−1.05, +1.05) for I_Mg, (−0.80, +0.80) for I_Fe, (−0.74, +0.74) for I_Cu, (−0.80, +0.80) for I_B, (−0.93, +0.93) for I_Zn, (−1.04, +1.04) for I_Mn, and (−1.03, +1.03) for the residual value. Compared to CND, the CVA approach shows weak detection of the nutrients that cause nutritional imbalance. CND indices revealed, except for N, the presence of nutrient imbalances and the necessity to correct the mineral nutrition of date palm in the Kebeli oases.

Despite its inconveniences, the most recognized method to extract boron from soils is that of hot water extraction (B_HW), which is used for diagnostics and recommendations. However, the Mehlich-3 (M3) method is widely used to extract and diagnose several elements at once (P, K, Ca, Mg, Al, B, Cu, Zn, Fe, and Mn) and is well adapted to routine analyses. The objective of our study was to develop a soil diagnostic and recommendation system for boron as a function of measured B_M3 (and other interacting elements), crop type, and spreading methods. This system is based on three databases from either the international literature or the chemical characterization of acidic-to-neutral soils typical from Québec (Canada). The first database came from the characterization of 365 samples typical of Québec soils; it has been used to predict, by the AutoML (Automatic Machine Learnig) supervised learning algorithm, B_M3 as a function of a set of parameters from the following: B_HW, pH_W, organic carbon (OC), Ca_M3, K_M3, and Mg_M3. Depending on the parameters used, the R² between the measured and observed B_M3 varied from 0.36 to 0.99. This database allowed us to define two classifications for soil boron diagnostics and fertility evaluation. The Cate–Nelson analysis for these two models allowed us to define three boron fertility classes: Low, medium and high; that is 0.00–0.23, 0.23–0.58, and 0.58–3.70 mg B kg⁻¹, respectively, for B_HW, and 0.00–0.65, 0.65–1.03, and 1.03–12.70 mg B kg⁻¹, respectively, for B_M3. The third database was extracted from 130 yield responses to increasing levels of boron; it was used to define a recommendation model for boron, based on AutoML, as a function of B_M3, pH_W, the crop boron requirement (medium, high), and the type of spreading (broadcast, sidedress, foliar spraying). This model resulted in an R² of 0.63.

Soil texture is a key soil property influencing many agronomic practices including fertilization and liming. Therefore, an accurate estimation of soil texture is essential for adopting sustainable soil management practices. In this study, we used different machine learning algorithms trained on vis–NIR spectra from existing soil spectral libraries (ICRAF and LUCAS) to predict soil textural fractions (sand–silt–clay %). In addition, we predicted the soil textural groups (G1: Fine, G2: Medium, and G3: Coarse) using routine chemical characteristics as auxiliary. With the ICRAF dataset, multilayer perceptron resulted in good predictions for sand and clay (R² = 0.78 and 0.85, respectively) and categorical boosting outperformed the other algorithms (random forest, extreme gradient boosting, linear regression) for silt prediction (R² = 0.81). For the LUCAS dataset, categorical boosting consistently showed a high performance for sand, silt, and clay predictions (R² = 0.79, 0.76, and 0.85, respectively). Furthermore, the soil texture groups (G1, G2, and G3) were classified using the light gradient boosted machine algorithm with a high accuracy (83% and 84% for ICRAF and LUCAS, respectively). These results, using spectral data, are very promising for rapid diagnosis of soil texture and group in order to adjust agricultural practices.