Prediction of Boron Adsorption by Field Samples of Diverse Textures

tent, on the other hand, have the potential to attenuate phytotoxic B concentrations (Goldberg, 1993). Detailed Soil texture often varies dramatically in both vertical and horizontal quantification of B adsorption reactions is necessary to directions in field situations and affects the amount of B adsorbed and B movement. This study was conducted to evaluate the effect of understand the fate and transport of B in soils. This clay content on B adsorption and to test the predictive ability of the knowledge is essential for determining appropriate appliconstant capacitance model to describe B adsorption as related to cations of B fertilizer, management of irrigation waters changes in clay content. Boron adsorption on 15 soil samples constitutcontaining large amounts of B, and reclamation of soils ing five depths of each of three sites in the western San Joaquin containing large amounts of B. Valley of California was investigated. Boron adsorption increased Boron adsorption on soil constituents has been dewith increasing pH, reached an adsorption maximum around pH 9, scribed by various empirical and chemical modeling apand decreased with further increases in pH. The model was able to proaches. The parameters obtained with empirical moddescribe B adsorption on the soils by simultaneously optimizing three els such as the distribution coefficient, Kd, and Langmuir surface complexation constants. The model was able to predict B and Freundlich adsorption isotherm equations are only adsorption by using surface complexation constants calculated from easily measured chemical parameters. The model was also able to valid for the particular conditions of the measurement predict B adsorption at all of the depths using the surface complexation (Goldberg, 1993). Chemical models, such as surface constants predicted with the chemical properties of one of the surface complexation models, account for molecular features depths and a surface area value calculated from clay content. These and define specific surface species, chemical reactions, results are very encouraging, suggesting that for a particular soil series, mass balances, and charge balances thermodynamically B adsorption for various sites and depths in a field can be predicted (Sposito, 1983). These characteristics allow chemical using only clay content and the chemical information from a different models to have more general predictive capability. This site in the same field. Incorporation of the prediction equations into has been demonstrated in prior studies for B adsorption chemical speciation-transport models will allow simulation of soil using the constant capacitance model, a surface comsolution B concentrations in horizontal and vertical space under diplexation model (Goldberg et al., 2000, 2004). verse environmental and agricultural conditions. A general regression model was developed to predict B adsorption on soils by calculating surface complexation constants for the constant capacitance model from B is a trace element essential for the growth of prediction equations (Goldberg et al., 2000). These prehigher plants, but the plant sufficiency range is diction equations relate the three surface complexation narrow (Reisenauer et al., 1973). In areas of plentiful constants: the B adsorption constant, KB , the protonrainfall, plant deficiency symptoms are often observed ation constant, K , and the dissociation constant, K , because of small soil solution B concentrations and large to the easily measured soil chemical parameters: surface amounts of B leaching (Keren and Bingham, 1985). In area, organic carbon (OC) content, inorganic carbon arid areas, B toxicity symptoms are primarily the result (IOC) content, and free aluminum oxide content. The of large soil solution B concentrations and application equations reliably predicted B adsorption envelopes of large amounts of B in irrigation waters (Nable et al., (amount of B adsorbed as a function of solution pH at 1997). Both B deficiency and toxicity conditions inhibit a fixed total B concentration) with reasonable accuracy plant growth leading to marked yield reductions of crop on 15 soils primarily from California (Goldberg et al., plants and economic losses to growers. 2000) and B adsorption isotherms (amount of B adSoil solution B concentrations equilibrate with B adsorbed as a function of equilibrium solution B concensorbed onto various organic and mineral surfaces (Goldtration) on 22 Midwestern soils (Goldberg et al., 2004). berg, 1993). Adsorption sites on organic matter, oxide These applications demonstrated a completely indepenminerals, clay minerals, and carbonates act as sources dent evaluation of the predictive capability of the conand sinks for B. Adsorbed B is neither directly available stant capacitance model to describe B adsorption by soils. nor toxic to plants (Keren et al., 1985). Thus, the adsorpThe amount of B adsorbed on soil constituents has tion complex plays a critical role in controlling soil solubeen correlated with soil clay content (Elrashidi and tion B concentrations. Boron deficiency often occurs on O’Connor, 1982). Previous research showed that the sandy soils because of their small B adsorption capacity regression model and prediction equations of Goldberg (Goldberg, 1993). Soils with large amounts of clay conet al. (2000) were able to satisfactorily describe B adsorption on soils of diverse textures in a field situation Contribution from USDA-ARS, George E. Brown Jr., Salinity Lab., 450 W. Big Springs Road, Riverside, CA 92507. Received 9 Nov. 2004. *Corresponding author ([email protected]). Abbreviations: CRM, coefficient of residual mass; IOC, inorganic carbon; M, mean difference; OC, organic carbon; OF, optimized fit; Published in Soil Sci. Soc. Am. J. 69:1379–1388 (2005). Soil Chemistry PE, depth-specific prediction; PEs9, prediction with parameters for Site 9 0to 30-cm depth; PEs33, prediction with parameters for Site doi:10.2136/sssaj2004.0354 © Soil Science Society of America 33 0to 30-cm depth; PEs49, prediction with parameters for Site 49 0to 30-cm depth; RMSE, root mean square error; SA, surface area. 677 S. Segoe Rd., Madison, WI 53711 USA