410
Maarel and Franklin 2012). As mentioned in the introduction (Sect. 15.1), the
choice of methods will depend on the research question, the site conditions, equipment, and personnel available, among other things. There are many good protocols
available (see Sect 15.1); familiarizing oneself with a couple of options and their
advantages and limitations and testing them under specific scenarios is generally
good practice.
15.3.1 Leaf-Level Spectroscopy
A typical setup for leaf-level spectroscopy consists of a spectrometer, light source,
fiber-optic cable, leaf clip or integrating sphere, and user interface. Leaf-level spectrometers can be classified into VNIR instruments, usually covering the visible to
the beginning of the near-infrared (NIR) portion of the electromagnetic spectrum (~
350–1000 nm), and full-range instruments, covering additional wavelengths in the
NIR and the shortwave-infrared (~350–2400 nm). Generally, VNIR instruments use
a silicon array detector, which does not require cooling, making VNIR instruments
relatively light and easy to carry. Full-range instruments use additional indium gallium arsenide (InGaAs) photodiodes, which require cooling, to detect the longer
wavelengths in the less energetic infrared part of the spectrum, making instruments
heavier and less stable. The conditions at the field site should be kept in mind when
choosing an instrument. If the spectrometer needs to be carried for longer times and
does not come with its own backpack, some extra effort is required to figure out a
good packing solution, especially for the fiber-optic cable, which can be easily damaged. It is good practice to check with the instrument companies if warranties are
still valid when instruments are transported without their shipping cases; additional
insurance might be worth considering.
A number of leaf clips are commercially available; some are easier for one person to handle and/or better suited to measure narrow leaves, such as conifer needles
and grass blades, than others, and there is room for design improvements (e.g.,
using 3D printers). Leaf clip measurements can be used to calculate reflectance (the
ratio of detected to incident light), which most instruments do internally, while measurements with integrating spheres can be used to calculate reflectance, transmittance, and absorptance. Leaf clip measurements are generally faster. Ideally, leaves
should cover the entire field of view of the sensor; special protocols are available for
narrow leaves (Noda et al. 2013). It is important to note that measurements with
different setups and among different instruments cannot be directly compared (Hovi
et al. 2017). One relatively laborious way to make measurements comparable (e.g.,
to include data from different instruments in one study) would be estimating empirical transfer functions. This requires measuring the same leaf samples with the
instrument setups to be compared. Empirical transfer functions can be estimated for
each wavelength and applied to transform measurements from one spectrometer
and setup to the other, given that measurement conditions are comparable between
model development and application.
A. K. Schweiger
Maarel and Franklin 2012). As mentioned in the introduction (Sect. 15.1), the
choice of methods will depend on the research question, the site conditions, equipment, and personnel available, among other things. There are many good protocols
available (see Sect 15.1); familiarizing oneself with a couple of options and their
advantages and limitations and testing them under specific scenarios is generally
good practice.
15.3.1 Leaf-Level Spectroscopy
A typical setup for leaf-level spectroscopy consists of a spectrometer, light source,
fiber-optic cable, leaf clip or integrating sphere, and user interface. Leaf-level spectrometers can be classified into VNIR instruments, usually covering the visible to
the beginning of the near-infrared (NIR) portion of the electromagnetic spectrum (~
350–1000 nm), and full-range instruments, covering additional wavelengths in the
NIR and the shortwave-infrared (~350–2400 nm). Generally, VNIR instruments use
a silicon array detector, which does not require cooling, making VNIR instruments
relatively light and easy to carry. Full-range instruments use additional indium gallium arsenide (InGaAs) photodiodes, which require cooling, to detect the longer
wavelengths in the less energetic infrared part of the spectrum, making instruments
heavier and less stable. The conditions at the field site should be kept in mind when
choosing an instrument. If the spectrometer needs to be carried for longer times and
does not come with its own backpack, some extra effort is required to figure out a
good packing solution, especially for the fiber-optic cable, which can be easily damaged. It is good practice to check with the instrument companies if warranties are
still valid when instruments are transported without their shipping cases; additional
insurance might be worth considering.
A number of leaf clips are commercially available; some are easier for one person to handle and/or better suited to measure narrow leaves, such as conifer needles
and grass blades, than others, and there is room for design improvements (e.g.,
using 3D printers). Leaf clip measurements can be used to calculate reflectance (the
ratio of detected to incident light), which most instruments do internally, while measurements with integrating spheres can be used to calculate reflectance, transmittance, and absorptance. Leaf clip measurements are generally faster. Ideally, leaves
should cover the entire field of view of the sensor; special protocols are available for
narrow leaves (Noda et al. 2013). It is important to note that measurements with
different setups and among different instruments cannot be directly compared (Hovi
et al. 2017). One relatively laborious way to make measurements comparable (e.g.,
to include data from different instruments in one study) would be estimating empirical transfer functions. This requires measuring the same leaf samples with the
instrument setups to be compared. Empirical transfer functions can be estimated for
each wavelength and applied to transform measurements from one spectrometer
and setup to the other, given that measurement conditions are comparable between
model development and application.
A. K. Schweiger