Corn Wastes and Peanut Shell as Growing Media for Production of Red Radish Plants in Soilless System

ABSTRACT

Peat is considered the conventional growing medium in most soilless culture systems. The high cost of peat and the urgent need of agricultural wastes recycling encouraged the scientists and soilless culture users to search for an alternative growth medium where optimal growth conditions are achieved and help in the safe disposal of wastes. In the current study, peanut shell (PS) and corn wastes (CW) were used as growing media in comparison to peat moss (PM). The tested organic wastes and peat moss were examined with sand at three mixing ratios (1:1 “M₁”, 1:2 “M₂” and 1:3 “M₃” raw material: washed sand, respectively). Red radish (Raphanus raphanistrum subsp. sativus) plant were cultivated in 5 kg pots filled with the instigated growing media. Most of the recoded growth parameters were found in PS and PM growing media, while the lowest ones were found in CW. PS medium contained available N and P higher by 141 and 29% above the PM media. Although the peat moss gave the highest values in the measured growth characteristics, its high price decreased the net profit. The highest net profit value was obtained from PSM₂ followed by PSM₁ and PSM₃. According to the obtained results, the characteristics of the growth media derived from peanut shell qualify them for use in the production of red radish in soilless culture systems.     

PHYSICOCHEMICAL CHARACTERIZATION OF HAZELNUT HUSK RESIDUES WITH DIFFERENT DECOMPOSITION DEGREES FOR SOILLESS GROWING MEDIA PREPARATION

In this study, the main physical, physiochemical, and chemical properties of hazelnut husk have been determined for the suitability as a container growing medium. Four substrates were prepared based upon decomposition degree: raw (H₁), medium decomposed (H₂, H₃), and strongly decomposed hazelnut husk samples (H₄), to determine the optimum growing medium parameters. Particle size was the key parameter to determine the physical properties, which was decreased with the increasing decomposition degrees. The main physical parameters were in the ranges for the ideal growing medium characteristics for H₄. Except for raw material (H₁), chemical characteristics such as pH, electrical conductivity (EC), carbon (C)/nitrogen (N), and nutrients were within acceptable ranges. The results indicated that hazelnut husk could be an alternative growing media component for containerized crops, if optimum particle size within the range of 0-2mm is obtained, for both aeration and water holding requirements.     

Excess trace metal effects on cotton: 5. Nickel and cadmium in solution culture

Abstract

Two cultivars of cotton (Gossypium spp.) were grown in solution culture in a glasshouse to determine phytotoxicity effects of excesses of Ni and Cd. Leaf yield was depressed 94% by 10‐⁴ M NiSO₄(with 198μg Ni/g leaf) in Acala SJ‐2 and 93% (with 167μg Ni/g) in Plma PS‐5. The Ni gradient was roots > stems > leaves in both cultivars. At 10^‐5 M, CdSO₄ gave more phytotoxicity than NiSO₄. The 10^‐4 M CdSO₄ resulted in about the same amount of phytotoxicity as did the Ni for both cultivars. The Pima PS‐5 plant parts, however, contained less Cd than did the Acala SJ‐2 at the highest Cd concentration. At 10^‐5 M CdSO₄ the reverse held in leaves and stems. Interactions held for both metals but the inverse effect between Cd and Mn was less pronounced than for other species. Many other interactions were present.     

Peat replacement in horticultural growth media: the adequacy of coir,paper sludge and biogas digestate as growth medium constituents for tomato (Solanum lycopersicum L.) and lettuce (Lactuca sativa L.)

Purpose: Due to environmental concerns, efforts are made to replace the use of peat in horticultural growth media by organic wastes. Four growth media were prepared with the purpose of achieving adequate physical and chemical properties for plant production. Materials and methods: Growth media prepared from mixtures of coir (C) and paper sludge (P), respectively, with two biogas digestates from food waste (D₁ and D₂), were tested. These mixtures, 20% D₁ or D₂?+?80% C or P (v/v), were evaluated as growth media for tomato (Solanum lycopersicum L.) and lettuce (Lactuca sativa L.). Results and conclusion: The growth media were all physically stable during the growing period, provided all the macronutrients and most of the micronutrients necessary for plant growth, adequate pH conditions, as well as an adequate electrical conductivity. The mixture of D₂ and P produced the highest biomass compared to a mineral fertilised peat (control), with a biomass production of 76% of the control for lettuce and 54% for tomato. Causes for the biomass reduction relative to the control may be related to ammonium toxicity effects, and/or limited plant-available water. The digestates, particularly D₁, seemed also to have a phytotoxic effect on the germination.     

Role of organic fractions on C decomposition and N mineralization of animal wastes in soil

The relative contributions of water-soluble, water-non-soluble, Van Soest-soluble, and neutral detergent fiber (NDF) fractions of pig slurry (PS), cattle slurry (CS), cattle farmyard manure (FYM), and composted cattle farmyard manure (CFYM) to the overall C and N mineralization of the raw wastes were studied by incubating treated soil for 107 days at 15°C under non-limiting N conditions. The C or N mineralization of soluble fractions was calculated from the difference between C or N mineralization of the raw and non-soluble fractions. The organic N content of raw wastes ranged from 15 to 32 mg N g⁻¹ dry matter and organic C to organic N ratio from 13 to 29. The water-soluble fraction (SOL_W) was close to 100 mg C g⁻¹ raw waste C for CS, FYM, and CFYM but reached 200 mg C g⁻¹ for PS. The Van Soest-soluble fraction (SOL_VS) was the main fraction for PS, CS, and CFYM (>500 mg C g⁻¹ raw waste C) but only 303 mg C g⁻¹ raw waste C for FYM. Both soluble and non-soluble fractions contributed to C decomposition of slurries, with half to more than half of the decomposed C derived from the degradation of soluble compounds. Most of the C decomposed from FYM was derived from the large NDF fraction, but the contribution from the water-soluble C to the decomposition was also significant. Carbon mineralization of CFYM was due to the degradation of the NDF fraction, whereas soluble C did not contribute. Amounts of N mineralized or immobilized by raw wastes and non-soluble fractions at the end of incubation were significantly correlated (P < 0.01) with their organic C to organic N ratio. The contribution of the Van Soest-soluble fraction to N mineralization varied greatly between the four wastes. Finally, large differences in the C degradability and N availability of the water and Van Soest-soluble fractions were demonstrated.     

Relationships between electrical conductivity and fertilizer levels in an azalea Liner growing medium

Abstract

Regression analyses were used to survey the effects of extractable NO₃‐N, P, K, Ca, Mg, and soil pH on electrical conductivity (EC) in a 1: 1: 1 (v/v/v) peat‐sand‐perlite azalea (Rhododendron sp.) growing medium amended with soluble 21–3–6 (N‐P‐K) or slow‐release 18–3–6 fertilizer sources. Simple linear correlations indicated a high degree of relationship between NO₃‐N, P, and K versus EC (r=0.95**, 0.79**, and 0.79**, respectively). Partial regression coefficients suggested that NO.‐N was the primary ionic species affecting EC. The correlations between EC versus P and K were apparently related to their intercorrelation with NO₃ ‐N, associated with the use of N‐P‐K fertilizers. These relationships were constant across fertilizer sources. The results demonstrate a potential for the use of soil EC in the estimation of NO₃‐N in soilless rooting media.     

A comparison of base media used for the incubation of plant residues in laboratory experiments

Abstract

Laboratory incubation experiments designed to measure potential rates of decomposition and N mineralization from plant residues typically use either sand or soil as a base medium for incorporation of the materials. Few studies have addressed the effects of base media on decomposition and N mineralization, or the interaction between media and incubated materials. This experiment compared the decomposition and net N mineralization rates of four plant residues incubated in the laboratory in both sand and soil base media. Initially, CO₂‐C evolution was greater from residues incubated in die soil, while net N mineralization was greater with the residues in sand. After 11 weeks, cumulative net N and C mineralized from residues incubated in either medium were more similar, but differences between media were still present. The form of N recovered differed between the media, with nitrification absent or occurring at a low rate in the sand. Differences in the rate and extent of decomposition and net N mineralization from plant materials were evident. These differences could influence the interpretation of results from this type of experiment.     

Effects of hardwood sawdust in potting media containing biosolids compost on plant growth,fertilizer needs,and nitrogen leaching

Abstract

Biosolids compost is used in media to grow potted plants. Nitrogen (N) in media leachate may contribute to nitrate (NO₃‐N) contamination of surface or ground water. Addition of sawdust to potting media containing biosolids compost will increase the carbon (C) to nitrogen ratio and could prevent N leaching without adversely affecting plant growth. A control medium containing 0% sawdust (v/v), 30% perlite, 50% municipal biosolids compost, and 20% sand was modified to contain either 10, 20, or 30% (v/v) fresh hardwood sawdust. The sawdust replaced either 1/3, 2/3, or all of the perlite in the control medium. Slow release fertilizer, slow plus quick release fertilizer, or no fertilizer was added to each of the four media to determine how the sawdust affected fertilizer needs. Coreopsis (Coreopsis grandiflora L.) and Rudbeckia (Rudbeckia hirta L. ’Goldstrum') were grown in pots for five months. Leachate was tested for NO₃‐N and ammonium N (NH₄‐N). Increasing amounts of sawdust produced no differences in growth of Coreopsis and few differences in the growth of Rudbeckia. The addition of slow or slow plus quick release fertilizer had little effect on the growth of Coreopsis and a greater effect on the growth of Rudbeckia. Sawdust and fertilizer had no effect on the leaching of N. Nitrogen leached primarily as NH₄‐N during the first four weeks of the experiment.     

Effects of Two Organic Wastes in Combination with Phosphorus on Growth and Chemical Composition of Spinach and Soil Properties

Abstract

Most agricultural soils in Iran are usually low in organic matter (OM). Therefore, increasing OM in these soils is of great concern. Environmental pollution caused by chemical fertilizers has created an interest in the integrated use of organic wastes with inorganic fertilizers. The main purpose of this greenhouse study was to evaluate the impact of two organic wastes and phosphorus (P) on the growth, and elemental composition of spinach (Spinacia oleracea L.) and soil chemical properties. Treatments consisted of four levels of municipal waste compost, MWC (0, 1, 2, and 4%), five rates of poultry manure, PM (0, 1, 2, 3, and 4%), and three P levels (0, 25, and 50 mg kg^?1 as KH₂PO₄). Application of P and MWC alone or in combination significantly increased the top dry weight of spinach. However, spinach growth was markedly increased up to 3% PM and suppressed with the higher rate, probably due to an excess of soluble salts in the soil. Moreover, the enhancing influence of P on spinach seedling growth was more pronounced at lower levels of MWC and PM. Plant P concentration tended to increase with increasing P, MWC, and PM application rates, whereas nitrogen (N) concentration was only affected by the two organic wastes treatment. Manganese (Mn) concentrations decreased, and copper (Cu), lead (Pb), and cadmium (Cd) increased by soil P application. However, P addition significantly decreased zinc (Zn) concentration only in MWC-treated spinach. Spinach plants enriched with either of the two biosolids accumulated more Mn, Zn, Pb, Cd, chloride (Cl), and sodium (Na) than control plants. Furthermore, spinach grown on MWC-amended soil contained higher Mn, Zn, Cu, and Pb and lower N, Cl, and Na than those raised on PM-treated soil. Postharvest soil sampling indicated that application of the two biosolids significantly increased concentration of soluble salts, (EC_e), OM, TN, NaHCO₃-extractable P, and DTPA-extractable iron (Fe), Mn, Zn, Cu, Pb, and Cd.     

Effect of pH and incubation temperature on nitrogen drawdown index of woodwaste potting media

Abstract

The effect of pH and incubation temperature on the nitrogen (N) drawdown index (NDI) of woodwaste potting media was investigated. A sand/pinebark/sawdust medium (1:1:8) was amended with seven rates of dolomite to give seven media with pH in the range 5.3 to 6.6. Nitrogen Drawdown Indices (NDI) for the addition of 75 mg N/L (NDI75) and 150 mg N/L (NDI₁₅₀) were measured for each medium after incubation at 10, 13, 20, and 30°C. NDI was unaffected by pH of the medium but was significantly influenced by incubation temperature. Nitrogen immobilisation was minimal at 10°C but increased with rising temperature such that at 30°C even the NDI₁₅₀ was zero. It is suggested that NDI tests be performed under conditions similar to those in which the potting media will be used in order to accurately assess potential N immobilisation under production conditions. A significant relationship between NDI assessed as NDI₇₅ and NDI₁₅₀ is reported. The relationship indicates that potting media with NDI₁₅₀>0.42 and NDI₁₅₀>0.81 will conform to the requirements of the Australian standard for regular and premium grade media respectively.     

EFFECT OF DIFFERENT PHOSPHORUS SOURCES ON THE GROWTH,YIELD, ENERGY CONTENT AND PHOSPHORUS UTILIZATION EFFICIENCY IN MAIZE AT RAWALAKOT AZAD JAMMU AND KASHMIR,PAKISTAN

Effect of poultry manure (PM) and four inorganic phosphorus (P) fertilizers sources, i.e., diammonium phosphate (DAP), single super phosphate (SSP), nitrophos (NP) and triple super phosphate (TSP) on crop production and P utilization efficiency (PUE) of maize was studied. Both inorganic P fertilizers and PM applied alone or combined in 50:50 proportions at equivalent rate of 90 kg P₂O₅ ha^?1. Results indicated that inorganic P sources with PM significantly increased plant height, leaf area and chlorophyll content. Average values showed that combined application of inorganic P with PM increased grain yield by 19 and 41% over inorganic P and PM alone, respectively. Similarly, increase in P-uptake due to the combined application of inorganic P + PM was 17% compared to sole inorganic P. Phosphorus utilization efficiency of inorganic P was increased with PM and the highest PUE was recorded in DAP + PM. Generally, combination of DAP + PM proved superior over the remaining P fertilizers.     

Assessment of iron availability in soilless potting media

Abstract

Ferrous sulphate added to potting media based on wood wastes is rapidly rendered insoluble in water. The extractability by DTPA of the added Fe declines over about 7 days to a steady value which is maintained for at least 10 months whether plants are growing in the media or not. The sawdusts and barks tested all required at least 100 mg/L Fe, added as ferrous sulphate, to provide an optimum amount of available Fe; peats needed less Fe and were quite variable in their requirements. In several pot trials, excellent correlations (R² 0.80–0.86) were obtained between quality or growth of shoots and a combination of medium pH and the concentration of Fe in unbuffered DTPA extractants of concentrations in the range 0.001 to 0.005M, or a saturation extract containing DTPA. Fe concentrations in DTPA/TEA/CaCl₂, NaHCO₃/DTPA, 0.5M ammonium acetate solutions and water were less precise indicators of Fe availability. At media pH values of 5.5–6.0. 0.002M DTPA (1:1.5 volume) extracts need to contain at least 20 mg/L Fe and preferably 30 mg/L Fe for adequate Fe supply if ferrous sulphate is the source of added Fe. Some species need at least 40 mg/L Fe or must be grown at a lower pH. These criteria do not hold if the main source of Fe is FeEDDHA.     

Evaluation of three calcium extractants for coastal plain soils

Abstract

Mehlich 1‐Ca is used as an index to predict the Ca requirement for peanut (Arachis hypogaea L.) fruit development in major peanut growing states. Recently, some concern has been raised about the inadequacy of Mehlich 1 extractable Ca for that purpose. Possible use of alternative extractants for soil Ca has been suggested. In this study, relationships among Mehlich 1, 0.2 M NH₄Cl and 0.01 M NaNO₃ extractable Ca were examined in several Coastal Plain soils to which gypsum or lime had been applied. Variability in extractable Ca was much greater following lime treatment than following gypsum treatment. In Bonifay soil, the quantity of Ca extractable by the three extractants was similar in a gypsum treatment, but in a lime treatment (at an application rate equivalent to the gypsum treatment) Mehlich 1‐Ca was 2 and 5‐fold greater than NH₄Cl‐ and NaNO₃‐Ca, respectively. In Greenville soil, Mehlich 1‐Ca was 3 to 4‐fold greater than NaNO₃‐Ca regardless of gypsum or lime amendment.

For soil samples from a field experiment on Lakeland sand, where lime or gypsum was applied prior to planting, Mehlich 1‐Ca was 7.5 and 2.2‐fold greater than NaNO₃‐Ca for the lime and gypsum treatments, respectively. Greater variability in Mehlich 1‐Ca in lime than in gypsum treatments was due to possible inclusion of undissolved limestone in the soil samples, resulting in overestimation of Ca available for peanut fruits. Mehlich 1‐Ca appears to be an adequate index of soil Ca for prediction of supplemental Ca requirement for peanut if lime has not been applied or has been applied well in advance of planting, thus minimizing the inclusion of undissolved limestone with the soil sample taken from the fruiting zone (0–8 cm depth) 10–14 d after planting.     

Effect of Zn binding to phytate and humic substances on its uptake by wheat (Triticum durum L.) as affected by carbonates and Fe oxides

Although complexation with soil organic matter may improve zinc (Zn) bioavailability to plants, the effect of Zn sorbent surface on the use of complexed Zn by plants remains unknown. The objective of this research was to elucidate how Zn complexation with humic substances (HS) and phytate affects the uptake of Zn by wheat plants depending on the main sorbent surface in growth media, i.e., carbonates and Fe oxides. To this end, two pot experiments were performed, one using Fe oxide-coated siliceous as the siliceous growth medium sand and the other using a mixture of calcareous sand and siliceous sand as the calcareous growth medium. Each experiment involved three Zn sources, Zn-HS complex, Zn phytate, and ZnSO₄. All sources were applied with surface irrigation at two Zn rates (0.25 and 2 mg kg^-1 growth medium). The Zn-HS complex significantly increased Zn uptake by plants in both media, relative to the other two Zn sources, but no significant difference was observed between Zn phytate and ZnSO₄. In the calcareous medium, Zn-HS complex and Zn phytate resulted in significantly higher dry biomass yields of wheat than ZnSO₄. In the siliceous medium, spike and shoot dry biomass yields with Zn-HS complex at the low rate and Zn phytate at both rates were not significantly different from those with ZnSO₄ at the high rate. After harvest, approximately 50% of the Zn applied as Zn-HS complex remained extractable by diethylenetriaminepentaacetic acid (DTPA), while this proportion was less than 20% for the other Zn sources. Thus, Zn-HS complex and Zn phytate are sources of available Zn for plants, and they are more effective than ZnSO₄ in increasing plant growth, particularly when carbonates are the main Zn sorbent surface.     

Losses of iron in leachates from organic components of potting media

Abstract

Losses over 18 weeks of iron (Fe) in leachates from several peats and wood wastes that had been amended with ferrous sulfate (FeSO₄‐7H₂O) were generally less than 1% of the total Fe in the material. Increasing additions of Fe either had little effect on Fe losses in leachates or lowered them, mainly through a lowering in the concentration of Fe‐organic complexes in leachates. DTPA‐extractable Fe in wood wastes did not decline with leaching but there were some reductions in peats. The data suggest that a single pre‐plant addition of 0.75–1.0 g/L FeSO₄‐7H₂O would supply enough Fe for at least six months of plant growth in media based on peat with a low native Fe content.     

Changes in the characteristics of potting media with time

Abstract

Growing media for container plants composed of pine bark, peat and sand were evaluated before, during and after five months with ryegrass as the test plant in the greenhouse. Hydraulic conductivity of these materials is very rapid resulting in rapid loss of nutrients, even for the controlled‐release type fertilizers. Channels developed on the perimeter of the growing medium ball adjacent to the container wall resulting in a concentration of plant roots in this zone. Considerable shrinkage or compaction of the media occurred during the five months growth period, especially for peat. Movement of sand was observed for the very porous materials and found to accumulate as a layer lower in the container. All nutrients added were exhausted after 3 months, even for the controlled‐release types of fertilizers.     

Acidification of lime‐stabilized biosolids in formulation of synthetic topsoil

Abstract

Uses of lime‐stabilized biosolids in container media or topsoil formulations may be limited by alkalinity imparted by lime added for reductions in vector attraction and pathogens and for suppression of odors. This experiment was conducted to assess the suitability of quicklime‐stabilized biosolids mixed with calcareous or noncalcareous sand and with sulfur or aluminum sulfate as acidifying agents for development of synthetic topsoil for turfgrass (Lolium perenne L.) production. Grass emergence and growth were inhibited in media that were not amended with an acidifying agent. Additions of aluminum sulfate were more effective in acidification than sulfur, causing an immediate drop in pH, whereas additions of sulfur required about 2 months for a depression in media pH. With time, electrical conductivities were higher in media with additions of sulfur than in media with additions of aluminum sulfate. Calcareous sand mixed with biosolids (2 sand: 1 biosolid, v: v) gave a better medium for growth of turfgrass than noncalcareous sand mixed with biosolids. Calcareous sand flocculated biosolids, whereas noncalcareous sand dispersed them. Results suggest that calcareous sand and aluminum sulfate are effective agents for diluting and acidifying biosolids for development of a suitable topsoil medium for turfgrass production.     

Effects of biochar and poultry manure on soil properties,growth, yield and quality of cocoyam (Xanthosoma sagittifolium Schott) grown in sandy soil

ABSTRACT

Field experiments were conducted during the 2017 and 2018 cropping seasons, to evaluate the effects of biochar (B) and poultry manure (PM) on soil physical and chemical properties, leaf nutrient concentrations, growth, mineral composition and corm and cormel yield of cocoyam. The experiment each year consisted of 4 × 2 factorial combinations of B (0, 10, 20 and 30 t ha^?1) and PM (0 and 7.5 t ha^?1). Results of the study indicated that in both years, the application of B and PM alone, and in combination, improved soil physical and chemical properties, leaf nutrient concentrations, growth, mineral composition and corm and cormel yield of cocoyam. There was a significant interaction effect of B and PM (B x PM) which was adduced to the ability of the B to increase PM-use efficiency and promote better use of the nutrients in the PM. It was found that combination of 30 t ha^?1 B and 7.5 t ha^?1 PM (B₃₀+ PM_7.5) gave the highest corm and cormel yield of cocoyam compared with other treatments. The combination of 30 t ha^?1 B and 7.5 t ha^?1 PM (B₃₀+ PM_7.5) exhibited the highest impact and is therefore recommended for soil sustainability and cocoyam productivity on sandy soil.     

Gypsum material effects on peanut and soil calcium

Abstract

Adequate availability of calcium (Ca) in the upper 7–10 cm of soil is extremely important for pod development and therefore for production of quality peanut (Arachis hypogaea L.). Supplemental Ca is usually applied as gypsum, however, availability of Ca may depend on the type of gypsum. The objective of a laboratory study was to evaluate recovery of Ca in Mehlich I, 0.01 M NaNO₃ and deionized water extractants from seven gypsum materials which varied physically from fine and coarse powders to crystals, granules and pellets. Overall, recovery of Ca was much greater in Mehlich I (89.5–99.6% of total Ca) than in either 0.01 M NaNO₃ (81.0–98.4%) or deionized water (78.7–97.5%). However, for 3 sources, recovery of Ca was very similar in NaNO₃ solution and deionized water.

Field experiments were conducted on Lakeland sand (Mehlich I Ca = 127 kg/ha, 0–15 cm) and Tifton loamy sand (Mehlich I Ca = 665 kg/ha) soils to study the effects of the gypsum materials on Florunner peanut grade and yield. Effects of gypsum treatments on Mehlich I‐ and 0.01 M NaNO₃‐extractable soil Ca were also evaluated during peanut pod development. On the Lakeland soil, Mehlich I Ca increased from 127 to a range of 420–737 kg/ha following application of gypsum depending on the type of gypsum material. Percent sound mature kernels were significantly greater where gypsum was applied than in the control treatment, regardless of source. Total sound mature kernel yield and gross return were greatest for the more soluble sources. The yield vs. soil test Ca relationship 90 d after planting revealed that yield response was very minimal if soil Ca was greater than 290 and 85 kg/ha of Mehlich I and 0,01 M NaNO₃‐extractable Ca, respectively. Application of gypsum to the Tifton soil, regardless of type of material, had no significant effect on yield, grade or gross return because Mehlich I extractable Ca in the control treatment was well over 560 kg/ha; the critical soil test Ca for runner peanut according to the current Georgia soil test recommendation.     

Application of VNIR and machine learning technologies to predict heavy metals in soil and pollution indices in mining areas

Purpose

Soil pollution indices are an effective tool in the computation of metal contamination in soil. They monitor soil quality and ensure future sustainability in agricultural systems. However, calculating a soil pollution index requires laboratory measurements of multiple soil heavy metals, which increases the cost and complexity of evaluating soil heavy metal pollution. Visible and near-infrared spectroscopy (VNIR, 350–2500 nm) has been widely used in predicting soil properties due to its advantages of a rapid analysis, non-destructiveness, and a low cost.

Methods

In this study, we evaluated the ability of the VNIR to predict soil heavy metals (As, Cu, Pb, Zn, and Cr) and two commonly used soil pollution indices (Nemerow integrated pollution index, NIPI; potential ecological risk index, RI). Three nonlinear machine learning techniques, including cubist regression tree (Cubist), Gaussian process regression (GPR), and support vector machine (SVM), were compared with partial least squares regression (PLSR) to determine the most suitable model for predicting the soil heavy metals and pollution indices.

Results

The results showed that the nonlinear machine learning models performed significantly better than the PLSR model in most cases. Overall, the SVM model showed a higher prediction accuracy and a stronger generalization for Zn (R²_V?=?0.95, RMSE_V?=?6.75 mg kg^?1), Cu (R²_V?=?0.95, RMSE_V?=?8.04 mg kg^?1), Cr (R²_V?=?0.90, RMSE_V?=?6.57 mg kg^?1), Pb (R²_V?=?0.86, RMSE_V?=?4.14 mg kg^?1), NIPI (R²_V?=?0.93, RMSE_V?=?0.31), and RI (R²_V?=?0.90, RMSE_V 3.88). In addition, the research results proved that the high prediction accuracy of the three heavy metal elements Cu, Pb, and Zn and their significant positive correlations with the soil pollution indices were the reason for the accurate prediction of NIPI and RI.

Conclusion

Using VNIR to obtain soil pollution indices quickly and accurately is of great significance for the comprehensive evaluation, prevention, and control of soil heavy metal pollution.