last update 26-Feb-2002
Abstracts of Contributions to the 2002 Ocean Sciences Meeting
13:30h AN: OS32J-01
TI: Non-electronic Sources of Biogeographical Data
AU: * Fautin, D G
EM: fautin@ku.edu
AF: Ecology and Evolutionary Biology, Natural History Museum, and Kansas Geological
Survey, University of Kansas, Lawrence, KS 66045 United States
AB: Most historical data and many data currently being collected that are relevant
to marine biogeography are unavailable electronically. Putting them into a form
that can be stored and used electronically is time-consuming but essential for
many purposes. Historical data provide a time dimension of centuries, producing
a baseline obtainable in no other way when environmental change is occurring
on a scale of decades. Even point measurements of environmental variables can
be informative. Taxonomic identification of very few kinds of organisms is possible
by remote sensing. Assembling information from museum catalogs -- even electronic
ones -- cannot produce comprehensive taxon lists except, perhaps, for taxa with
few members. The presumed difficulties of capturing non-electronic data are
primarily those of entry. The human effort involved in entering these data is
not so different from that needed to manipulate electronic data (by converting,
editing, parsing, etc.) to make them useful for particular purposes.
UR: http://www.kgs.ukans.edu/Hexacoral/
DE: 1694 Instruments and techniques
DE: 4299 General or miscellaneous
SC: OS
MN: 2002 Ocean Sciences Meeting
1330h AN: OS42C-140
TI: Ocean-Scale Biogeography: Predicted Distributions of Anemonefish Sea
Anemones
AU: * Baker, J
EM: bayjaker@hotmail.com
AF: Department of Zoology, Brigham Young University, Provo, UT 84602-5255 United
States AU: Sandhei, P
EM: psandhei@kgs.ukans.edu
AF: Department of Geography and Kansas Geological Survey, University of Kansas,
Lawrence, KS 66045 United States
AU: Fautin, D G
EM: fautin@ku.edu
AF: Dept. Ecology and Evolutionary Biology, Natural History Museum, and Kansas
Geological Survey, University of Kansas, Lawrence, KS 66045 United States
AB: Only tens to a few hundreds of georeferenced occurrence data are available
in electronic form for the sea anemones that host the obligately symbiotic anemonefishes,
although these animals occur in tropical waters from the eastern shores of Africa
and the Red Sea to French Polynesia and Japan. We investigated whether the environmental
characteristics of places where the anemones are known to occur could be used
to predict accurately where they are not known to exist but do live. We obtained
known localities of the anemones and environmental data from the Hexacorallia
database, and used the geospatial clustering tool LOICZView to identify the
environmental parameters that define suitable habitat for the anemones (all
are at www.kgs.ukans.edu/Hexacoral). Initial tests were done using unsupervised
clustering of the environmental variables mean depth, mean monthly sea surface
temperature (SST), mean salinity, and wave height. Although promising, the results
were less generalizable than desired. In refining the clustering, the best results
were from mean depth after excluding minimum depths greater than 100 m, minimum
monthly SST, minimum monthly salinity, wave height, ocean color, tidal range,
and coral reef occurrence. In addition to more selective definition of environmental
parameters, known occurrences of anemones were used to train the prediction
process. The revised clusters were tested for ability to predict occurrence
of anemonefish, which served as indicators of anemone occurrence. Our preliminary
results indicate that 1) reef occurrence is a good predictor of anemones and
the fishes that live with them, 2) environmental clusters supervised with data
on anemone occurrence are equally good predictors of anemonefish occurrence,
and 3) we were readily able to identify about one-third of the potential range
that has occurrence probabilities substantially better than random chance.
UR: http://www.kgs.ukans.edu/Hexacoral/
DE: 4899 General or miscellaneous
DE: 1694 Instruments and techniques
DE: 4299 General or miscellaneous
SC: OS
MN: 2002 Ocean Sciences Meeting
1330h AN: OS42C-139
TI: Evaluating standards for digital representation of locality information
AU: * Hunsinger, K L
EM: hunsing@ku.edu
AF: University of Kansas, 1475 Jayhawk Blvd, Lawrence, KS 66045-7613 United
States
AB: Locality data can be used to link environmental and biological data, but
digitally representing localities reported in scientific literature is complicated
by variation in areal precision and in how they are expressed. I describe three
ways to georeference a locality digitally: using a single point to represent
a locality; using polygon(s) for exact areal definition of the locality; and
identifying bins (grid cells) that correspond to the locality. I evaluate these
approaches using the criteria of accuracy, transportability, and cost. An accurate
digital representation should describe each locality completely and specifically,
quantify its areal precision, and indicate the source of information. Transportable
data are easily moved into and compatible with other databases, which provides
flexibility, especially for future studies. The cost of time and materials of
data input should be minimized. Given a report that a marine animal was collected
somewhere in Australia, no single point can adequately represent the possible
locations of that site, and areal precision cannot be objectively quantified.
Defining a polygon that includes all the area that might be called an Australian
marine environment, and excluding all area that would not, is very accurate
and allows easy quantification of areal precision, but requires much time and
effort. Creating a grid and identifying all the bins that represent an Australian
marine environment is more accurate than using a single point, allows quantification
of areal precession, and is easier than defining a polygon(s). A grid of latitude
and longitude lines is easy to make but the bins are not equal in size, so the
precision with which a record can be captured will vary according to latitude.
Equal-area grids do exist, but they are difficult to make and compare to existing
databases. Supported by NSF grants DEB-9521819 and DEB-9978106 to Daphne G.
Fautin (in the program Partnerships to Enhance Expertise in Taxonomy), and OCE-0003970
to DGF and R. W. Buddemeier (in the National Oceanographic Partnership Program).
UR: http://www.kgs.ukans.edu/Hexacoral/Biodata/index.html
DE: 4255 Numerical modeling
SC: OS
MN: 2002 Ocean Sciences Meeting
15:45h AN: OS32J-06
TI: Land Forcing and Coral Reefs: Terrestrial Runoff as a Factor in Coral Reef
Distribution
AU: * McLaughlin, C
EM: caseym@kgs.ku.edu
AF: University of Kansas, Dept. Of Geography 213 Lindley Hall, Lawrence, ks
66045 United States
AU: Smith, C
EM: csmith3@swarthmore.edu
AF: Swarthmore College, Dept. of Engineering, Swarthmore, PA 19081
AB: Coastal ecosystems such as coral reefs are increasingly in danger from non-local
anthropogenic effects such as deforestation, land use, and pollution in inland
river basins. These non-local pressures are channeled from a potentially large
basin scale through freshwater discharge into the coastal zone. As a first estimate
of a reef-to-runoff relationship, we examined global reef distributions as a
function of total runoff within a 30' grid cell. We interpret the resulting
correlation as meaning that runoff inhibited reef occurrence when runoff was
greater than $10^10$ m3/yr. Combining basin runoff and five additional variables
(average sea surface temperature, minimum salinity, wave height, tidal range,
Chlorophyll-A) selected to proxy the effect of runoff, increased predictive
capabilities. The use of statistical representation of spatial and temporal
variability allowed useful analytical comparisons of the environmental variables.
Spatial and temporal summary statistics (mean, standard deviation, extremes)
were summarized for each variable into a standard 30' spatial grid cell, providing
a common framework for K-means clustering routine. Classification of runoff-related
stresses were then extended, for example, by adding modeled sediment discharge
to refine the prediction of areas of reef stress from human activities. Information
on such environmental controls is important to understanding both paleo-environmental
forcing of reefs and the potential effects of present and future human alterations
to the hydrologic cycle.
DE: 0910 Data processing
DE: 1630 Impact phenomena
SC: OS
MN: 2002 Ocean Sciences Meeting
11:15h AN: OS41M-10
TI: The Concept or the Number: Problems of Scale, Precision, Visualization,
and Communication
AU: * Buddemeier, R W
EM: buddrw@ku.edu
AF: Kansas Geological Survey, University of Kansas 1930 Constant Avenue, Lawrence,
KS 66047 United States
AU: Maxwell, B A
EM: maxwell@swarthmore.edu
AF: Swarthmore college, Swarthmore College, Swarthmore, PA 19081 United States
AU: Bartley, J D
EM: jbartley@kgs.ukans.edu
AF: Kansas Geological Survey, University of Kansas 1930 Constant Avenue, Lawrence,
KS 66047 United States
AB: Increasingly, information can move in an automated fashion from sensing
device to database to analytical tool to final product, often with an admixture
of other data and substantial transformation or processing along the way. Disciplinary
conceptual models of such processes are based implicitly on assumptions that
everybody agrees on and knows the appropriate pathway and form and presentation
of the product. These assumptions are not valid in cross-disciplinary applications,
where checkpoints and alternative pathways in data flow and processing are critical.
Shared visualization (in 2, 3, 4, or even more dimensions) is vital to scientific
cooperation and communication, but raises the geographer's dilemma: the most
rigorous or scientifically accurate representation is often not the most subjectively
informative. Particularly when different types of variables (intensive/extensive,
classified/continuous, skewed/normally distributed) are combined in a single
analysis or model, mismatches in units, data handling, or transformations may
compromise the desired results. For example, unit conversions alter apparent
precision, and differences between latitude-longitude and equal-area grid systems
are immaterial for normalized variables (concentrations or surface densities),
but can be critically important if quantitative budgets are desired. The use
of global-scale environmental data sets in conjunction with local-scale biological,
ecological, and biogeochemical data has provided numerous opportunities to experience,
and occasionally to address, the need to retain human participation in automated
data management and application processes. We will present illustrative examples
and suggest guidelines for appropriate types and levels of data automation -
and non-automation - for various kinds of applications.
UR: http://www.kgs.ukans.edu/Hexacoral/
DE: 4899 General or miscellaneous
DE: 1699 General or miscellaneous
SC: OS
MN: 2002 Ocean Sciences Meeting
09:35h AN: OS11S-05 INVITED
TI: Reefs as Habitats or Habitats for Reefs: Global-Scale Coral Reef Biogeography
AU: * Buddemeier, R W
EM: buddrw@ku.edu
AF: Kansas Geological Survey, 1930 Constant Avenue, Lawrence, KS 66047 United
States
AU: McLauglin, C J
EM: caseym@kgs.ukans.edu
AF: Kansas Geological Survey, 1930 Constant Avenue, Lawrence, KS 66047 United
States
AU: Sandhei, P
EM: psandhei@kgs.ukans.edu
AF: Kansas Geological Survey, 1930 Constant Avenue, Lawrence, KS 66047 United
States
AB: Coral reef organisms tend to exhibit very wide geographic distributions.
Reefs in the structural sense are necessarily features with histories of centuries
to millennia. This large-scale perspective is difficult to incorporate into
consideration of present concerns about the death and degradation of reef communities,
at local sites and on time scales of days to decades. We suggest that a critical
approach is to regard coral reef communities as assemblages of organisms having
habitat requirements that transcend the concept of the reef community itself
as a unique habitat. This approach permits definition of habitat for reef organisms
that extends beyond existing reefs, and allows a broader and more integrated
definition of suitable -- or vulnerable - regions that may play a role in the
preservation of reef biota. Although global-scale analyses are constrained by
the availability and resolution of global-scale data sets, geospatial statistical
approaches combined with geographic information systems (GIS) can yield useful
insights into controls over reef organism distribution and how these may change
over time. We apply geospatial clustering in combination with GIS analysis to
identifying potential reef habitat based on existing reef distributions, and
to classifying clusters of habitats based on their probable stability. Although
not a substitute for detailed local assessments, such analyses provide a basis
for incorporating larger-scale considerations of biogeographic controls and
environmental change into reef habitat assessments at the national or local
scale.
DE: 1630 Impact phenomena
DE: 1699 General or miscellaneous
DE: 4804 Benthic processes/benthos
SC: OS
MN: 2002 Ocean Sciences Meeting
1330h AN: OS42C-138
TI: Data, Data Everywhere - And Not a Way to Choose
AU: * Misgna, G
EM: gmisgna@kgs.ukans.edu
AF: Kansas Geological Survey, University of Kansas 1930 Constant Avenue, Lawrence,
KS 66047 United States
AU: Bartley, J D
EM: jbartley@kgs.ukans.edu
AF: Kansas Geological Survey, University of Kansas 1930 Constant Avenue, Lawrence,
KS 66047 United States
AU: Buddemeier, R W
EM: buddrw@ku.edu
AF: Kansas Geological Survey, University of Kansas 1930 Constant Avenue, Lawrence,
KS 66047 United States
AB: The proliferation of electronic data sets has been a mixed blessing for
researchers and others interested in environmental and biological information.
The information potential in the expanding data offerings is often obscured
or unrealized by the inability of would-be users to make effective evaluations
and comparisons of data sets for the purposes of interest to them. Even within
fairly well-defined user communities and applications, data selection can be
a tedious and uncertain process. In the projects supported by the joint database
of the Land-Ocean Interactions in the Coastal Zone (LOICZ) and the Biogeography
of the Hexacorallia projects, the intention is to make useful, relevant data
broadly available to non-specialist and multidisciplinary user communities.
This requires standard formats for visualization and presentation, a convenient
means of reviewing the variables and datasets available, ready access to both
local and primary metadata, and importantly, means of visualizing both the numerical
and the geographic distributions of the data within a given set. These needs
have been addressed by adopting a grid system with appropriate scale and classifications,
and by constructing a dynamic `front end' for a web-served relational database.
This design provides rapid access and flexible development. This presentation
describes not only the underlying structures, but also some of the tools provided
as part of the data selection and download process. These permit the user to
select geographic or numerical ranges, filter or transform the data, exclude
or modify selected ranges of values, view single-variable distributions as histograms
or scatter plots, and construct correlation matrices for multiple variables.
For a relatively modest investment of development time, these features greatly
increase both the use, and the appropriateness of the use, of the data.
UR: http://www.kgs.ukans.edu/Hexacoral/
DE: 4899 General or miscellaneous
DE: 1699 General or miscellaneous
SC: OS
MN: 2002 Ocean Sciences Meeting
1330h AN: OS42C-141
TI: Environmental GIS Modeling of Distribution Patterns in Actinodendron
plumosum, a Sea Anemone With a Large Geographic Range.
AU: * Ardelean, A
EM: adorian@ku.edu
AF: University of Kansas, Division of Biological Sciences Haworth Hall, 3002
1200 Sunnyside Ave., Lawrence, KS 66045 United States
AB: I use locality records to plot the distribution pattern of morphotypes as
a way to test the hypothesis that several named species of the sea anemone genus
{\it Actinodendron} actually comprise a single species, {\it A. plumosum}. GIS
tools, prediction algorithms such as LOICZView and GARP, and existing environmental
databases can be used not only to predict distribution patterns but also to
solve taxonomic problems in marine biota with large geographic range. The known
distribution of these sea anemones consists of sparse data points with various
grades of precision. The associated environmental parameters can be used to
predict the geographic range of each morphotype. I comparatively analyze the
predicted distribution patterns to test a species hypothesis. Overlap between
distributions of morphotypes supports the hypothesis of synonymy. Geographical
separation of morphotypes can be used as evidence that the morphotypes belong
to different species.
UR: http://www.kgs.ukans.edu/Hexacoral/
DE: 4800 OCEANOGRAPHY: BIOLOGICAL AND CHEMICAL
SC: OS
MN: 2002 Ocean Sciences Meeting