1) Spectral response patterns embody considerable information about the composition and concentration of materials, and their acquisition is inexpensive and rapid. As such, the technique is profoundly useful for improving our basic understanding of landscape patterns (heterogeneity) and performance. For example, where accurate chemometrics can be developed, high resolution maps of plant, soil, or litter chemistry can be developed. Further, where significant spatial and temporal variability would otherwise confound observational studies, the increased statistical power made possible by high sample throughput methods makes control of confounders and elucidation of controlled fixed effects possible.
2) Evaluation of ecosystem condition requires synthesizing multiple measures that can be used as proximate indicators. For example, we often use soil phosphorus levels as a measure of performance (in agronomic circumstances) or degradation (along cultural eutrophication gradients). However, ecological condition is a multi-state, multi-stressor problem that is not always revealed using bi-variate signals. In response, efforts to identify proximate measures of performance are multi-variate, coupling the conditional effect of many measured variables. Where spectroscopy may have a role is in the routine assessment of ecosystem condition – because spectra embody information that may be used to predict many of the proximate indicators of condition, my research is examining the ability to develop direct predictions of condition from spectra, largely obviating the need for measurement of proximate variables. There are numerous examples of integrated spectral assessment in the literature.
The array of environmental samples for which rapid assessment technology could useful is enormous. Demonstrations in the literature for plant tissues, soils, wood, animal tissues, mineral samples, water samples, sediment samples have been successful, but rely on the development of comprehensive spectral libraries for training the statistical interpretations that are place and material specific.
The basic theory of spectroscopy derives from the behavior of dipole bonds under incident light of various frequencies. As with any oscillating system, certain frequencies cause resonance in the molecular bonds, and consequently absorbance. Beer’s Law suggests that the intensity of the optical effects (absorbance, reflectance, transmittance) that a sample has on incident light is directly proportional to the concentration of constituent bonds. As such, the location and magnitude of optical features in a spectrograph (for example, a graph of reflectance vs. wavelength) can be used to identify the constituent parts.
Soils are comprised of organic material (decaying plant and animal tissues), mineral components, pore spaces and water. Each constituent in
