GLOSOLAN-Spectroscopy webinar | Propagation of spectral measurement errors in mid-infrared spectroscopy | 25th June 1:00pm - 2:30 pm CET
With the increasing demand for high-resolution regional, continental and global soil data to monitor the current state of the soil, soil spectroscopy has been increasingly used to obtain measure soil properties as addition to, or full replacement of, conventional analytical methods. To obtain quantitative predictions of a soil property, wavenumbers, of which the spectrum consists, are calibrated against reference analytical soil data. Once calibrated on the wavenumbers, the components of which a spectrum consists, the resulting spectral model can be used to predict the soil property of interest solely by new spectral information. Besides analytical measurement errors in the reference analytical soil data, a topic addressed in the previous webinar, measurement can also occur in the spectra. Instrumental drift or imprecise preparation of the microplates, can cause measurement error in the spectra. In this webinar, measurement error in mid-infrared (MIR) spectra will be discussed. Based on a case-study, the effect of measurement error in MIR spectra on model performance of spectral models will be demonstrated.
Speaker: Dr. Cynthia van Leeuwen
Biography: Dr. Cynthia van Leeuwen is a postdoctoral researcher at the Soil Geography and Landscape group of Wageningen University. She works for the ReGeNL project, a project aiming to transform the Dutch agricultural sector into a regenerative, sustainable, and future-proof industry. She uses remote sensing techniques to support monitoring and modelling efforts in the project. Previously, she completed her PhD thesis, titled ‘Statistical modelling of analytical and spectral soil measurement errors’, in 2024 at ISRIC and Wageningen University. In this thesis, analytical and spectral measurement errors in soil data were quantified, emphasizing the importance of informing data users about these errors before calibrating and validating models. Measurement errors propagate through soil models and can drastically reduce model performance and prediction uncertainty. In this thesis, the sensitivity of widely employed models within the soil science community to analytical and spectral measurement errors was studied.