ARS | Publication request: ASSESSING THE ACCURACY OF THE LINEARIZED LANGMUIR MODEL

Submitted to: American Society of Agronomy Meetings
Publication Type: Abstract Only
Publication Acceptance Date: July 28, 2006
Publication Date: October 14, 2006
Citation: Bolster, C.H. 2006. Assessing the accuracy of the linearized langmuir model. American Society of Agronomy Meetings.

Technical Abstract: One of the most commonly used models for describing phosphorus (P) sorption to soils is the nonlinear Langmuir model. To avoid the difficulties in fitting the nonlinear Langmuir equation to sorption data, linearized versions are commonly used. Although concerns have been raised in the past regarding the use of linearized equations, this practice is still commonly used today for describing P sorption to soils. The goal of this research was to investigate more fully the impact of linearization of the Langmuir equation on the accuracy of fitted parameter values and assessments of goodness of fit. Three different linearized versions of the Langmuir model were fit to sorption data collected on three western Kentucky soils and fitted parameters and goodness of fit measures were compared with results from the nonlinear equation. Both linear and nonlinear equations were also fit to two sets of synthetic sorption data: one with constant measurement error and the other with measurement error proportional to the measured value. Results show that the use of the most commonly used Langmuir linearization can give erroneously high r2 values due to inherent self-correlation caused by data transformation. Numerical simulations showed that the most accurate Langmuir equation will depend on the error structure of the measurements. For constant measurement error the nonlinear equation provided the best fits whereas when error is proportional to the measurement, a linearized version of the Langmuir equation will yield the best fits. Results of this study will allow researchers to make more informed decisions when applying the Langmuir model to their sorption data.