June 14

Contents

Editor's Introduction

This week, I report on the use of GIS for archaeology and for the new Atlas of Yellowstone. Plus, sixteen press releases.

— Matteo Luccio

GIS for Archaeology

Beyond 2D and 3D, GIS can also be used to map in 4D! The fourth dimension, of course, is time, and the application archaeology: start with a surface survey, then dig to add stratigraphic control and peel back layers of time. That is how Shinu Abraham, an assistant professor in the Anthropology Department at Saint Lawrence University, is planning to ultimately use the GIS that she is beginning to set up to organize and analyze ceramic fragments at an archaeological site. Because she is not trained in GIS, she partnered with a colleague who is, Carol Cady, also on the university's faculty. Ultimately, Abraham hopes to add GIS to her archaeological skill set and use it routinely in her work.

Abraham, an archaeologist who specializes in South Asian archaeology, has been focusing on the Southern part of India. Her larger interest is the Indo-Roman trade — the ancient maritime trade that took place between the various regions along the Indian Ocean littoral and, in particular, between Roman Egypt and South India.

Abraham has been working on a recently discovered site called Pattanam, which was discovered a few years ago on the southwestern coast of India, in the state of Kerala. "The reason this is interesting," Abraham says, "is that this is the earliest urban settlement that we have discovered so far in this region of India. It is particularly interesting to us because we realize that it is a port site that was involved, we think fairly heavily, in this inter-regional trade. We know about it from early historical sources — from South Indian texts from about 2,000 years ago and ancient Greco-Roman texts — and we think that it may be the fabled site known as Muziris, which was discussed a great deal in the early texts." Finding the physical remains of this site, Abraham explains, is a very exciting discovery and it allows archaeologists to excavate it and survey it in detail, in order to understand exactly how it was structured and organized, and how it integrated South Indian society with this international trade system.

Abraham's part of the project is a detailed surface survey of the site — collecting materials that exist on the surface in order to understand how the site was set up and whether there are satellite settlements or burial sites associated with this urban center. To do that, she started the Malabar Region Archaeological Survey (MRAS). For the past two archaeological seasons she has taken students to the site and begun to survey it systematically — mapping it and collecting surface artifacts. Ninety percent of the artifacts that Abraham and her team have collected are ceramics. Now she wants to analyze them, classify them, and examine their distribution across the site.

"It soon became clear to me," Abraham says, "that there was no way that I was going to be able to handle the wealth and diversity of data that I was going to be accumulating without some type of management facility and GIS was the obvious way to go." St. Lawrence has been trying very hard, she says, to build up the use of GIS as a tool across the different disciplines.

I asked Abraham a few questions about this project.

1. What area does the site cover?

   That is one thing that we have to determine. I and other scholars have done a preliminary reconnaissance of the site and we estimate that it is about 24 hectars, but that could change. We need to do the survey to establish exactly what the boundaries of the site are and how it filters into the surrounding landscape.

2. Have you been using aerial photography?

   The closest we've been able to do is use satellite imagery. One of our constraints is that this is a coastal site and it is very near some Indian naval bases. Therefore, there are several restrictions. We are limited to whatever datasets we are able to access. So, we have our satellite imagery, we have decades-old topographical maps that we are using, and we are also collecting a lot of existing published data about sites that are in and around the area in which Pattanam is located, so that we have some type of context for it.

3. What reference system do you use to survey the site?

   We take a GPS receiver into the field with us. One of the constraints of the site is that it is not in open space; it is sitting underneath a contemporary village. That means that the surface of the site is occupied by a series of small households within the village. So, rather than creating an arbitrary grid system and then using that as my dividing units, we are taking each of these very small households and using them as artificial units. Then, within each household, we are collecting all the surface remains. What we end up having, then, is a series of small plots that are laid out across the map. What I hope to do after I collect all of those smaller units is to group them into a series of grid cells, which I can then use to understand the layout of the site, the density in terms of the distribution of ceramics. As the ceramics peter out, we assume that that will help us define the extent of the site. Then, as we move beyond the site, into the outer part of the area, [we hope to] be able to find out whether there are satellite settlements associated with it.

4. At what stage of the project are you?

   I am hoping that this will end up being a five-year survey project, of which I've done two seasons of survey. We've collected data for the eastern third of the site, about 120 collection units all together. We hope over the next season and a half to collect more from within the site. By the end we hope to have 100 percent collection of the site itself and then a sample of the area immediately in the perimeter of the site.

5. Are you digging at all?

   There are restraints in terms of getting permits to actually do excavations. I'm doing a surface survey of the site and then I have colleagues who are doing excavations. I hope to eventually be able to incorporate their excavation data into the work that I am doing. The biggest value, of course, of excavating is that it will allow us stratigraphic control — it will give us an opportunity to understand the actual phases of occupation of the site. It will allow me to understand and evaluate the ceramics that I am collecting on the surface, because they will probably vary according to the phases of occupation. So, for example, we assume that the ceramics that are associated with the earliest phase — the megalithic or Iron Age phase —will encompass a much smaller area of the site. Then, as the site expanded into what is known as the early historic period, which is the time of the Indo-Roman and inter-regional trade system, we expect it to have gotten a little bit larger and a little bit more established. That is one of the things that we are hoping to demonstrate or at least test with this survey.

6. So, you can gather a lot of data without even disturbing the ground.

   Even though most people who are not archaeologists think that excavation is the only way to gather important and meaningful data about a site, work that has been done over the last several decades has shown us that survey work can actually be equally informative and much less invasive, especially when we are talking about private property. So, one of the benefits of GIS in this context is that it allows us to glean that much more detailed information without actually having to put a trowel into the ground.

7. How are you storing the data to put into the GIS?

   I just began doing my ceramic analysis last summer and creating a Microsoft Access database with all of the various attributes and the various types of ceramics. Then we will be able to introduce that data into the GIS and look at the distributions of different characteristics of ceramics across the site — for example, the distribution of bowls vs. jars, of different ceramic styles, and various other characteristics, such as temper and clay, rim measurements, etc. [This will help us distinguish] ceramics that were locally processed from those that were not.

8. What time period are you studying?

   I am particularly interested in the early historic period. Very roughly, this is from about 200 BC to 300 AD. There are some signs, from other research projects, that indicate that the upper limit of the early historic period may in fact go into the fourth and fifth century and that is something that we are hoping to be able to determine through the ceramic analysis as well.

9. When you go back this far, what are you finding on the surface?

   Lots and lots of ceramics, believe it or not! It is not uncommon for a site that has been occupied over a fair period of time. Even though I am interested in the early historic period, there are indications that there was an earlier phase, known as the Iron Age, which goes back to about 500 or 600 BC if not earlier, and there are phases after the early historic that include the medieval period, which goes on into the eighth and ninth centuries. So, we are seeing a whole range of materials, but all of those ceramics are, to some extent, represented on the surface. We are also lucky enough to find beads and also other indicators that it was an urban settlement — for example brick alignments and formations and ring wells (terracotta wells that were built into the ground). So there are many other features that help us speak about the site.

I asked Cady a few questions, too.

1. What is your role?

   I am providing GIS support, both actual support and moral support too. Part of this is how Shinu is actually learning GIS and will be able to use it in other projects.

2. What are your goals for this project?

   We want to use the data organizational capabilities of GIS to help organize, maintain, and display all the spatial data that she collects in the field. Because we are working in India, we don't have many data layers available at a very fine scale for what we need. We do have DEMs and some coarse level digital chart of the world data. We are trying to integrate other data sources, such as satellite imagery, to provide some of those base layers of information, so that she can go from the site level to integrate into a watershed level to look at trade routes.

3. Where did you get your satellite imagery?

   From Digital Globe. We had to pay for it, but there's an educational discount, so it makes it pretty reasonable. It is something like $500 per scene and we have about 8 scenes. We got it from their library, as opposed to having the images taken for us. It is sub-meter for the panchromatic. We're able to identify many of the structures, the roads, the canals, the types of trees, so that we can do a great deal of analysis with those photos to develop our base layers. We are in the process of doing some manipulation of the satellite imagery in Erdas Imagine, before we even bring it into the geodatabase, like doing mosaiking and that kind of thing.

4. What GPS receiver is Abraham using?

   She is using a Trimble GeoXM with 1-3 meter accuracy. In the village in which she is working right now, they are using GPS to map the roads and identify structures (e.g. where a temple is located).

5. What software are you using?

   She is storing her data in an Access table. We will be porting it into the geodatabase that we are developing. We are working with ArcGIS. We've been doing this in 9.1 and we are just migrating to 9.2. Right now we are storing everything as a personal geodatabase. This is one of the first geodatabases I've been working with on our campus. This is probably our big move into geodatabases. We will be storing all her ceramic data, all the GPS data, the satellite imagery, and any other data sources we may come across. Some of the data were originally in shapefiles and we are also migrating that into the geodatabase.

6. How is she collecting the data in the field?

   She has a laptop that has ArcGIS on it. She has very disjunct field seasons, so mostly the GPS data is collected in the field and then all the post-processing is done when she comes back. A technician who does GPS and GIS support is working with her on the post-processing.

The Atlas of Yellowstone

The largest display in the map gallery at the Twenty-Seventh Annual ESRI International User Conference in San Diego will be a set of 57 8 ft. x 3 ft. posters displaying a work in progress: the Atlas of Yellowstone: A Celebration of the Natural Diversity, People, and History of Yellowstone, by the University of Oregon's InfoGraphics Lab. This display encompasses half of the National Park Exhibit. Mid-way through a planned four-year course, the project uses GIS both to analyze data collected by specialists in a variety of disciplines, as well as to display it, in the same style as the Lab's landmark Atlas of Oregon. Additionally, the publication will also illustrate the process used to collect, analyze, and display the data.

"Some of the posters are showing off page pairs for the atlas and other ones talk about the process of how we use GIS to integrate data into Adobe Illustrator and Natural Scene Designer to produce these pages," says Alethea Steingisser, the project's cartographer. "Yellowstone also has a very rich on-line digital slide library, so we were able to use many of these historic images as graphic elements in the posters as well. We have, maybe, a quarter of the pages nearing final." Each poster tells about either the content of the atlas—for example, the relationship of American Indians to Yellowstone or the impact of tourism on the park—or the process of making the maps. Steingisser uploaded the files for the posters to an ESRI server and the company printed them at its Redlands headquarters and shipped them to San Diego for the show.

Each page of the atlas supports one or more of its core themes — such as how Yellowstone Park is connected to the greater Yellowstone area and how that area, in turn, is connected to the bigger region; the human imprint on the landscape; and the importance of Yellowstone as the first national park and as the epicenter of grizzly bear conservation efforts. In addition to the hundreds of maps developed specifically for the atlas, the publication will also contain many historic photos and maps, some drawn by the first white people to explore the area.

The first 30 or so posters for the gallery exhibit show how natural processes impacted Yellowstone, while the balance show how experts — such as wildlife biologists and geologists — collected the data. One poster, for example, shows how wildlife biologists collect data on the movements of animals using radio collars, another how a thermal inventory of geysers was done, and another how ethnographic data is collected. Other posters, Steingisser explains, show "the process of how we go from taking an idea" — for example, the park's complicated relationship with grizzly bears — "and turning that into page pairs for the Atlas." (Upon seeing a poster that included a sketch for an atlas page to explain the atlas' production process, I jokingly asked Steingisser whether another poster displayed the process used to produce the previous poster, and so on…)

One poster deals with some of the workflow issues: how the team gathers content for the atlas, reviews the data it receives, begins to pre-design the maps and the atlas pages, and then does the final page layout. "We start making a preliminary map," Steingisser says, "bring it into Adobe Illustrator, Photoshop, and Natural Scene Designer and do some more advanced design, then we put it in the mock-up. At this stage we put together our initial graphic and send it back to the park and talk to people whose research it depicts and ask them, 'Does this tell the right stories?' So, it is constantly a circular process."

"They don't just hand us data," Steingisser continues, "and say 'Go to it.' Usually we sit down with someone who studies, for example, fire, and talk with them about their research. What are the key issues? What should we be showing? Then we work with them to find out what data is available, what we can utilize. It is a longer process than the park just giving us a bunch of data."

During a large forest fire in 1988, researchers collected fire data and made daily polygons of the burned areas. The atlas team used those polygons and created fields to extract data out of that data set. It then used a color gradient to show the fire's progression, from when it started to when it was completely out.

Wildlife biologists have studied extensively the park's wolf packs — as old ones dissolve new ones form and different packs' territories overlap — often using collars equipped with GPS receivers and radio transmitters. The atlas team has used this data to map wolf populations using kernel density analysis and Hawth's analysis tool, a plug-in for ArcGIS that is commonly used by wildlife biologists. (The kernel density analysis outputs a raster, based on the points input by the user. Hawth's outputs shapefiles, based on the percentage of points that a user wants a polygon to include.) On this topic, even more than other ones, the team works very closely in conjunction with specialists in the park. "We are not wolf biologists," says Steingisser. "People that know the data know the stories. They let us know what's important about the data and interpret it."

In some cases, the atlas itself has inspired new ways of bringing data together. For example, the atlas' main map of American Indians shows their trails and what different tribes called different places. According to Steingisser, that information has never been brought together before. Likewise, the team developed a new way of mapping bison ranges and sent the resulting map to Congress, where it was used in meetings about bison management.

The amount of analysis required differs from data set to data set. "Sometimes," Steingisser explains, "they give us datasets that are very rough — say, 10,000 or 20,000 data points — and we just sit down and start playing around and think of what maps we can make that would tell a significant story. For example, they gave us 10,000 points of thermal features in the park, which is just a small fraction of them, and they also gave us polygons for thermal areas. On that one we did a quantitative analysis [to determine the density of geysers]. Other times, such as in the case of the wolf population, we will get data that's already been manipulated in some way by researchers who have been studying an issue for a very long time" — such as kernel density analysis to show the ranges and movements of wolf packs.

How does the team deal with incomplete data sets? "Our motto, from the Atlas of Oregon, is to map what data is available," says Steingisser. "So we don't necessarily sit and wait for data that is being collected right now. We are dealing with datasets that have already been collected and are already available." However, she points out, shortly before the atlas is published, the team will have to update the older data.

To display the maps created in GIS, the team uses Adobe Illustrator and other Adobe products. "I don't export any text out of GIS," Steingisser explains. "I do all my text and label placements and color choices in Adobe. Once we get the data that we want to show and select out the attributes that we want to show and do any type of analysis, we use it to get the vector linework out into Adobe." To output shaded relief in 2D and to create 3D perspective views, she first uses Natural Scene Designer, then brings the images into Photoshop to process them further.

A few atlas panels illustrate the initial meetings that took place at Yellowstone a couple of years ago, when Jim Meachan, the lab's director and the atlas' cartographic editor, Dr. Andrew Marcus, the atlas' lead editor, and Ann Rodman, Yellowstone Park's GIS coordinator and the atlas' Yellowstone editor, met with various subject-matter experts and discussed what the main story was for each topic. They took extensive notes and did graphic mock-ups, which now guide the layout of each atlas page. However, Steingisser points out, the team is constantly revising those layouts as it works with the data. This week, Steingisser is at Yellowstone for another round of meetings with the subject-matter experts to pick new topics for the atlas and to figure out how best to tell the story of the park and the area.

While the park has so far provided most of the data for the atlas, it comes, in turn, from many different sources, such as the U.S. Geological Survey (USGS), people collecting data for a thermal inventory, and historic geyser information from journals and electronic monitoring. When the atlas is finally published, this metadata will be in the back, in the form of references.

Working with Marcus and Meacham, Steingisser wrote her master's thesis on the human impact on geysers, focusing specifically on Yellowstone. At the same time, as a graduate teaching fellow, one of her duties was to work on the atlas. After she graduated, Meacham hired her as a staff cartographer and researcher, working primarily on this project but also on the Cultural Atlas of the Mongolian Altai. Rodman, who has collaborated with Marcus for many years, has been the atlas project's main Yellowstone contact.

Currently, the project is only about 20 percent funded — with seed funding from Canon, the Yellowstone Park Foundation, and the University of Oregon — so Steingisser is the project's only full-time cartographer. She hopes that the poster exhibit at the ESRI conference will help generate financial support for the project. "Once we get more funding and can have more people working on it, we can have more going on at one time," she says.

» Back to our GIS Monitor JUN 2007 Issue