487AI-GEOSTATS: Plant specimen point data: call for help and/or collaboration
- Jan 5, 2002Dear Geostats Members:
I am a botanist at The New York Botanical Garden. My colleagues and I study
plants, and many of us focus on tropical plant groups. We have six million
specimens in our collection from all over the world. The emphasis of my
e-mail message to you today involves plants of tropical America
Two of us here at the garden have been applying multivariate statistical
procedures in our research to test patterns of species distribution and
morphological variation. We use these stats to test such things as species
boundaries, and to figure out which morphological characters are
contributing to patterns of variation. We use both quantitative and
qualitative data (i.e., leaf length in cm; presence/absence of leaf
pubescence-- ordinal data-- 0/1/2).
On the side, separate from stats, we use ArcView GIS to plot plant
distributions onto geospatial data layers. Currently I plot specimen
collection locality points into my ArcView projects. I then bring in
layers, such as USGS DEMs and other environmental data from ArcAtlas and
ArcWorld to allow more detailed visualization, presentation, and
understanding of plant distribution patterns.
We are now interested in bringing statistics and GIS together in order to
statistically analyze distribution patterns of plants and display the
results of this analysis in the GIS presentation system.
For instance, we have a growing database for our plants. The database
represents information collected from individual plant specimens, including
locality, coordinates, habitat, and plant morphology.
How can we bring regression, correlation, multivariate statistics, and more
into our GIS work?
Where does one run the calculations? Can we do that in ArcView? We have
ArcView Spatial Analyst, but it seems to me that calculations are very
simple with that extension, i.e., nothing as powerful as a statistics
package such as SysStat or others.
Another challenge we have is multiple specimens per locality. Imagine that
each specimen is represented by the same 50 morphological data points
(measurements and ordinal scores). If there are 50 specimens at one point
(i.e., Iquitos, Peru), and each of the specimens has 50 data points, when we
run a simple grid analysis of one morphological character (i.e., leaf
length), the resulting grid point system represents only the specimens
sitting 'on top'. Do you understand? This is one example of problems we
Is anybody interested in a collaboration that would lead to collaborative
grants and projects? We're thinking of setting up a model system (or using
a system that has already been developed) to analyze distribution of our
specific plant groups (palms, wild nutmeg trees, and Brazil nut trees--
several thousand species in total). We would want to correlate
morphological trends with geospatial data to better understand distribution
Currently we are looking at collections from Mesoamerica (Mex and Central
America). By simple plotting of points we are already seeing patterns that
have never been published. There are gaps in distribution that relate to
historical deforestation, minute and major changes in habitat, changes in
rainfall, etc. These patterns have significant implications for
conservation of Mesoamerican habitats and for understanding distribution of
species, speciation patterns, and diversity patterns.
We are botanists. That is what we do best. We are not computer experts,
although they run the show these days in our work. If we spend all of our
time doing the stats and the geospatial analysis, we'll never have time to
continue generating the datasets for the plants we study. The datasets are
built through field and herbarium (botanical museum) work. We're also
discussing the idea of incorporating molecular data into the 'geospatial
system', especially population genetics data, but that is another story for
What we need is a collaborative team. Before we can build that team, we need
some feedback. Perhaps there are folks out there who are looking for
additional exciting projects from the computer and geospatial side.
If so, please contact us. We have six million specimens in the herbarium of
The New York Botanical Garden (www.nybg.org/bsci.htm). The specimens are
important for understanding species, for identifying species, and for
vouchering other studies based on these collections (e.g., molecular
analyses are vouchered by the herbarium specimen of the species being
analyzed). What we don't see is an effort to take these specimens and make
geospatial sense out of them.
We are interested in setting up a model system that will include:
1. plant specimens
2. plant species
3. geospatial data
4. gis mapping
5. geospatial analysis
8. presentation through paper publication
9. presentation through an interactive website at nybg.org
For the plants, we have plenty of data and some preliminary products. WE
will post some of the preliminary products to the internet this month. But,
our preliminary products do not take into account the amazing level of
resources and analysis that I think are available in the geospatial
If this sounds interesting, please contact us through me at:
Institute of Systematic Botany
The New York Botanical Garden
Bronx, NY 10458
(An e-mail contact is best to start with).
Thank you for your time! I am anxious to see if this e-mail will help me
generate some new contacts. I just signed up with this mailing list.
* To post a message to the list, send it to ai-geostats@...
* As a general service to the users, please remember to post a summary of any useful responses to your questions.
* To unsubscribe, send an email to majordomo@... with no subject and "unsubscribe ai-geostats" followed by "end" on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list
* Support to the list is provided at http://www.ai-geostats.org
- Next post in topic >>