409
15.3 Field Data Collection
Spectra of plants can be acquired across scales (see Gamon et al., Chap. 16), including at the leaf level, using proximal RS techniques (e.g., handheld spectrometers,
robotic systems, UASs), airborne instruments, and satellite systems. For leaf-level
studies, it is often of interest to collect information about taxonomic identity (species or clade), functional type (e.g., based on life form, growth form, dispersal type),
functional traits (e.g., based on samples for chemical or structural analysis, growth
measurements), developmental stage, and stress symptoms (e.g., signs of disease,
herbivory, drought). For canopy-level studies, it is common to collect information
about community composition and cover, spatial arrangement (or clustering), gap
fractions, plant and canopy architecture (e.g., leaf area index, leaf angle distribution,
branching structure, stem diameter, stratification), community biomass, and community traits. Additional data often collected together with vegetation spectra
include soil characteristics (e.g., chemistry, water content), elevation, slope, and
aspect. Important metadata include time and precise location, observer, nomenclature used, and photos, from which, for example, cover fractions can be estimated.
Ideally data are recorded digitally to avoid the time and sources of error associated
with transcriptions, and it is good practice to develop and test protocols for standardized data collection. Information should always be recorded as precisely as
possible. For example, in grasslands it would be unnecessary to record vegetation
height in classes because recording vegetation height at the cm level takes about the
same amount of time, and classes can always be aggregated later if needed. Working
together with other research groups can make it more efficient to collect additional
data. This requires coordination at an early planning stage.
Offering educational opportunities might be part of the mission of a research
area, and site administrators might be able to help with hiring students or technicians. However, it is advisable to focus on collecting the most ecologically relevant
data, using well-trained personnel and sound methods, including appropriate sampling design and large enough sampling size, rather than collecting various kinds of
data of poorer quality. For studies with a RS element, it is important to acquire
accurate and precise coordinates of research plots and/or individuals to match their
locations to the image data. Triangulation can be used to estimate plot coordinates
from ground control points, and relative positions of individuals within plots can be
estimated from plot coordinates. The level of accuracy and precision needed depends
on the spatial resolution of the imagery, but professional surveying equipment can
be needed. Again, early planning is important, because finding rental equipment can
become difficult during peak season. Purchasing insurance for expensive equipment
might be advisable. Research areas might have periodic surveying campaigns.
Including research plots in such campaigns is a great option but requires marking
plots temporarily; posts made out of a light but rot-resistant wood (e.g., larch,
spruce) are well suited for this.
The following sections give some examples about spectral data acquisition at
different levels of measurement; for details on the collection of ecological, nonspectral data, see textbooks on ecological methods (e.g., Sala et al. 2000; van der
15 Spectral Field Campaigns: Planning and Data Collection
15.3 Field Data Collection
Spectra of plants can be acquired across scales (see Gamon et al., Chap. 16), including at the leaf level, using proximal RS techniques (e.g., handheld spectrometers,
robotic systems, UASs), airborne instruments, and satellite systems. For leaf-level
studies, it is often of interest to collect information about taxonomic identity (species or clade), functional type (e.g., based on life form, growth form, dispersal type),
functional traits (e.g., based on samples for chemical or structural analysis, growth
measurements), developmental stage, and stress symptoms (e.g., signs of disease,
herbivory, drought). For canopy-level studies, it is common to collect information
about community composition and cover, spatial arrangement (or clustering), gap
fractions, plant and canopy architecture (e.g., leaf area index, leaf angle distribution,
branching structure, stem diameter, stratification), community biomass, and community traits. Additional data often collected together with vegetation spectra
include soil characteristics (e.g., chemistry, water content), elevation, slope, and
aspect. Important metadata include time and precise location, observer, nomenclature used, and photos, from which, for example, cover fractions can be estimated.
Ideally data are recorded digitally to avoid the time and sources of error associated
with transcriptions, and it is good practice to develop and test protocols for standardized data collection. Information should always be recorded as precisely as
possible. For example, in grasslands it would be unnecessary to record vegetation
height in classes because recording vegetation height at the cm level takes about the
same amount of time, and classes can always be aggregated later if needed. Working
together with other research groups can make it more efficient to collect additional
data. This requires coordination at an early planning stage.
Offering educational opportunities might be part of the mission of a research
area, and site administrators might be able to help with hiring students or technicians. However, it is advisable to focus on collecting the most ecologically relevant
data, using well-trained personnel and sound methods, including appropriate sampling design and large enough sampling size, rather than collecting various kinds of
data of poorer quality. For studies with a RS element, it is important to acquire
accurate and precise coordinates of research plots and/or individuals to match their
locations to the image data. Triangulation can be used to estimate plot coordinates
from ground control points, and relative positions of individuals within plots can be
estimated from plot coordinates. The level of accuracy and precision needed depends
on the spatial resolution of the imagery, but professional surveying equipment can
be needed. Again, early planning is important, because finding rental equipment can
become difficult during peak season. Purchasing insurance for expensive equipment
might be advisable. Research areas might have periodic surveying campaigns.
Including research plots in such campaigns is a great option but requires marking
plots temporarily; posts made out of a light but rot-resistant wood (e.g., larch,
spruce) are well suited for this.
The following sections give some examples about spectral data acquisition at
different levels of measurement; for details on the collection of ecological, nonspectral data, see textbooks on ecological methods (e.g., Sala et al. 2000; van der
15 Spectral Field Campaigns: Planning and Data Collection