Calibrating Spanner psychrometers for the effects of ambient temperature: theoretical and experimental considerations

Abstract

Calibration models are indispensable for the use of Spanner-type psychrometers, but the existing calibration models have not been adequately evaluated. A non-zero intercept in the regression between electrical output (E)and water potential (Ψ)is commonly observed during Spanner psychrometer calibration, but is incompatible with the temperature correction factor (TCF)model, used to describe the temperature-dependence of the E/Ψ relation. Calibration errors for 10 commercial, in-situ, temperature-compensated psychrometers, were compared over a temperature and Ψ range of 10–40 °Cand −2.5 to 0 MPa, respectively, using the current commercial TCF model, five modified TCF models, two empirical models, and a mechanistic model. TCF models differed in whether the E/Ψ relation was only determined at 25 °C and extrapolated to other temperatures, or whether the relation was established across temperatures, and whether model parameters were determined for each psychrometer (individual-fit)or all psychrometers (group-fit). Measurement error (deviation from linearity in the E/Ψ relation at each temperature), was 0.042 MPa, and only one individual-fit modified TCF model exhibited an error (0.048 MPa)which was not statistically different from measurement error. Individual-fit errors for other models ranged from 0.058 to 0.100 MPa. Group-fit errors for the Rawlins and all of the TCF models covered a relatively narrow range (0.067–0.070 MPa)compared to group-fit errors for all the empirical models (0.820–0.109 MPa).