5
Data Acquisition
Thomas J. Stohlgren
5.1 Introduction
The science of data acquisition has changed a lot
in the past two decades (Oppenheimer et al., 1974;
Michener, 1986). In 1974, Austin (in Oppenheimer
et al., 1974) marveled at the ability to view data
and computer outputs directly on a television
screen (cathode ray tube). It was a time when the
acronym GIS referred to "general information system" (Oppenheimer et al., 1974, p. 232). In 1986,
Klopsch and Stafford recommended the use of
8-inch single-sided, single-density or 5 1 /z-inch
double-sided, double-density diskettes for the storage of medium-sized data sets for the Long-Term
Ecological Research (LTER) program. Kriging was
touted as the best interpolation method for spatial
analysis (Seilkop, 1986). There is little doubt that
10 or 20 years from now our current commonly used
hardware (e.g., Unix-based workstations, Pentiumtype personal computers, digitizing tablets), storage media (e.g., CD-ROM, 3 1 /z-inch high-density
diskettes, tape backups), and software (e.g., ARC/
INFO geographic information systems, cokriging
spatial analysis programs) will invoke a similar
comedic response. However, some things haven't
changed; there is a consistent, unrelenting commitment to improve data acquisition and management in the ecological sciences.
The challenges in the art and science of data acquisition are growing. Changing ecological paradigms, issues, threats, and increasing demands by
society for accountability and management responsiveness encourage flexibility on the one hand;
on the other, metadata standards, data transfer protocols, and demands for comparable information
for multiple uses encourage increased standardization and rigidity. Data are expected to meet the
needs of individual studies, along with other local,
regional, national, and international needs, despite
meager local funding (Stohlgren et al., 1995b). The
push for interdisciplinary research involves more
investigators and the simultaneous collection of
data from multiple biological groups and environmental factors (Stohlgren et al., 1998). Investigations of ecological processes such as disturbance,
competition, herbivory, nutrient cycling, and energy flux involve integrated data sets from the laboratory, field experiments, observational studies,
and mathematical models. Spatially explicit data
and geographic information system (GIS) analyses
and modeling are expected outputs from many
studies (Stohlgren et al., 1997c). Computing capabilities are increasing at a linear rate relative to
data acquisition's exponential growth rate. Qualitychecked data are expected to be "loaded on the Web"
and accessible for public use in record-breaking time.
There is a perceived urgency for ecological assessments (see Chapter 1). Yet existing biotic inventories of natural landscapes commonly are poor
(Stohlgren and Quinn, 1992; Stohlgren et aI.,
1995a, 1995c). The paradigms of ecosystem management and adaptive management are founded on
the promise of a constant delivery of sound ecological data. Increased threats to ecosystems bring
increased demands for reliable data and increased
needs for ecological assessments.
The newest confounding factor to data acquisition is the issue of scale. The scale of inventory,
monitoring, and research programs has changed
from site by site assessments to the evaluation of
landscape and regional changes in resource sustainability, land-use patterns, and biodiversity (National Research Council, 1990, 1994). It is becoming increasingly important to extrapolate plot-level
information to landscape and regional scales (Scott
71
Data Acquisition
Thomas J. Stohlgren
5.1 Introduction
The science of data acquisition has changed a lot
in the past two decades (Oppenheimer et al., 1974;
Michener, 1986). In 1974, Austin (in Oppenheimer
et al., 1974) marveled at the ability to view data
and computer outputs directly on a television
screen (cathode ray tube). It was a time when the
acronym GIS referred to "general information system" (Oppenheimer et al., 1974, p. 232). In 1986,
Klopsch and Stafford recommended the use of
8-inch single-sided, single-density or 5 1 /z-inch
double-sided, double-density diskettes for the storage of medium-sized data sets for the Long-Term
Ecological Research (LTER) program. Kriging was
touted as the best interpolation method for spatial
analysis (Seilkop, 1986). There is little doubt that
10 or 20 years from now our current commonly used
hardware (e.g., Unix-based workstations, Pentiumtype personal computers, digitizing tablets), storage media (e.g., CD-ROM, 3 1 /z-inch high-density
diskettes, tape backups), and software (e.g., ARC/
INFO geographic information systems, cokriging
spatial analysis programs) will invoke a similar
comedic response. However, some things haven't
changed; there is a consistent, unrelenting commitment to improve data acquisition and management in the ecological sciences.
The challenges in the art and science of data acquisition are growing. Changing ecological paradigms, issues, threats, and increasing demands by
society for accountability and management responsiveness encourage flexibility on the one hand;
on the other, metadata standards, data transfer protocols, and demands for comparable information
for multiple uses encourage increased standardization and rigidity. Data are expected to meet the
needs of individual studies, along with other local,
regional, national, and international needs, despite
meager local funding (Stohlgren et al., 1995b). The
push for interdisciplinary research involves more
investigators and the simultaneous collection of
data from multiple biological groups and environmental factors (Stohlgren et al., 1998). Investigations of ecological processes such as disturbance,
competition, herbivory, nutrient cycling, and energy flux involve integrated data sets from the laboratory, field experiments, observational studies,
and mathematical models. Spatially explicit data
and geographic information system (GIS) analyses
and modeling are expected outputs from many
studies (Stohlgren et al., 1997c). Computing capabilities are increasing at a linear rate relative to
data acquisition's exponential growth rate. Qualitychecked data are expected to be "loaded on the Web"
and accessible for public use in record-breaking time.
There is a perceived urgency for ecological assessments (see Chapter 1). Yet existing biotic inventories of natural landscapes commonly are poor
(Stohlgren and Quinn, 1992; Stohlgren et aI.,
1995a, 1995c). The paradigms of ecosystem management and adaptive management are founded on
the promise of a constant delivery of sound ecological data. Increased threats to ecosystems bring
increased demands for reliable data and increased
needs for ecological assessments.
The newest confounding factor to data acquisition is the issue of scale. The scale of inventory,
monitoring, and research programs has changed
from site by site assessments to the evaluation of
landscape and regional changes in resource sustainability, land-use patterns, and biodiversity (National Research Council, 1990, 1994). It is becoming increasingly important to extrapolate plot-level
information to landscape and regional scales (Scott
71