| Official title of the project: | "Remote Sensing-based Geostatistical Modeling for Coniferous Forest Inventory and Characterization" | ||
| Investigators: | Dr. Mark Jakubauskas, Research Assistant Professor, KARS Program | ||
| Dr. Edward A. Martinko, Director, KARS Program | |||
| Dr. Kevin P. Price, Associate Director, KARS Program | |||
| What is the KARS Program ? | The Kansas Applied Remote Sensing (KARS) Program is a research program of the Kansas Biological Survey at the University of Kansas. KARS conducts research on environmental and agricultural applications of remote sensing technology. | ||
| Who funded the project ? | The project was funded by the National Aeronautics and Space Administration (NASA), Earth Science Enterprise. | ||
| Are other agencies involved ? | In 1998, a Memorandum of Understanding (MOU) was issued by Secretary Dan Glickman of the U.S. Department of Agriculture (USDA) and Mr. Dan Goldin, Administrator of NASA. The purpose of the MOU was to provide the framework to enhance the cooperative activities between NASA and USDA in the areas of joint science and applications research, and technology transfer concerning agriculture and related disciplines. | ||
| How was it selected ? | In 1998, NASA published a Request for Proposals under the heading of Remote Sensing Applications in Agriculture, Forestry, and Rangeland Research. One hundred eighty (180) proposals were submitted by scientists and companies from across the country; 13 were funded, including the KU proposal. | ||
| How long is the grant for ? | The grant is for three years. Results of the competition were announced by NASA in April 1999, and the project officially started August 1, 1999. | ||
| How much is the grant for ? | NASA awarded the KU project $560,000 total over the three years of the project. | ||
| What are geostatistics ? | Geostatistics are an advanced geographic analysis technique that take into account the characteristics of the area surrounding a point of interest. Simple statistics - such as correlation and regression - have been used in remote sensing for many years to describe the relationship between the amount of light reflected by an point or location, and some property or attribute of that area, whether that be vegetation biomass or soil moisture or water depth. By calibrating the satellite data with field data, we can use these statistical relationships to estimate properties of the land surface in areas we have not field-sampled. Geostatistics extends this by taking into account the properties of the points surrounding the point of interest to make more precise and accurate estimates. | ||
| What is remote sensing ? |
Remote sensing systems - of which aerial photography is but
one type - record the amount of light reflected or emitted by
an object, in specific bands of the electromagnetic spectrum.
Different land cover types reflect light in different ways; plants
absorb blue and red visible light (and therefore look green to
us) and reflect near-infrared light very strongly. Water, in
contrast, reflects some visible (blue, green, and red) light,
but almost completely absorbs near-infrared light. A photograph is a record of reflected visible light on film or paper. The remote sensing systems we use are far more advanced; they record light in parts of the spectrum we cannot see with the human eye; they record the information digitally, as a set of numbers, rather than film or paper; and they are carried on board earth-orbiting satellites, at about 400 miles up. |
||
| How much detail can you see in the imagery ? | Current remote sensing systems in orbit can detect objects as from about a half mile in size (1000 meters) down to as small as three feet across (one meter). We call this the spatial resolution. Most data we work with have a spatial resolution of about 30-100 feet (10-30 meters), although we will be testing data from several different systems to see which performs best in our models. | ||
| How large an area does a satellite scanner record ? | The satellite systems we most typically use can scan an area about 100 x 100 miles (180x180 kilometers), in about 30 seconds. This means we can have an image of the entire forest captured at the same time. Most air photos only capture an area about 3x3 miles per photo, by comparison. | ||
| Why Yellowstone ? |
Several reasons. First, we have been conducting research on remote sensing of the Yellowstone forests for nearly a decade, and know the area well. Jakubauskas' dissertation (KU, Geography, 1994) was on remote sensing of Yellowstone forests, and he has continued his work there since. Second, Yellowstone is an excellent area for developing the advanced methods for the project - the forests are relatively untouched, and are dominated by one species, lodgepole pine. The region is fairly flat, and permission to sample has been granted by the Park Service. |
||
| Are you doing field work ? | Yes. We completed the first summer of fieldwork on the project in 1999, gathering data on 330 field points. We will return during the summer of 2000 to add another 200-300 points. Finally, in the third and last year (2001), we will shift to a National Forest outside of Yellowstone. | ||
| Why the shift to the National Forest ? | This is a rigorous test of the ability of our system to accurately map the forests based on satellite imagery. We are using the data from Yellowstone to create and calibrate the geostatistical models. We will then generate predictions for the National Forest, and then field sample to see if our predictions were correct. | ||
| What KU students are involved in the project ? |
Todd Searls is an undergraduate in Biological Sciences who has participated in both the fieldwork and the satellite image processing. An outstanding graduate student from the University of Calgary, Monika Moscal, is starting her PhD in Geography this fall and working as a graduate research assistant on the project. Matt Ramspott assisted with the fieldwork in summer 2000, and is a Self Fellow starting his PhD in the Department of Geography that fall. |
||
| Who would use or could benefit from the results of the project ? | The project is aimed at forest managers and planners, who need both basic information on forest for inventory purposes (height, density, age, basal area, leaf area) and thematic information (wildfire risk, insect infestation risk, changes in the forest over time) for management and planning purposes. The system will be usable by both public foresters (in National Parks and Forests) and private companies that raise timber for paper pulp and lumber. | ||
| What products or services will result from the project ? | Although we can ourselves produce maps of forest characteristics or risk, the principal product will be a software system that foresters can easily use; satellite images and field-sampled data on the forest are the inputs, and the forest maps are the output. | ||
| Will the software be sold or commercialized ? | The software system will be commercialized via Terrametrics, Inc, a private company formed in cooperation with the University of Kansas to commercialize remote sensing research and applications developed by the KARS Program. In cooperation with TerraMetrics, KARS facilitates technology transfer of products and services derived from remote sensing technologies to commercial, governmental, and other end users. | ||
| Who are the key people at NASA to contact regarding the project ? |
Alexander (Alex) J. Tuyahov Rodney McKellip |
||
| Some relevant websites: | |||
| KARS Program homepage: | http://www.kars.ku.edu/ | ||
| KARS Forest Project website: | http://www.kars.ku.edu/forest | ||
| NASA Food and Fiber page : | http://www.crsp.ssc.nasa.gov/ffars/ffarsmain.htm | ||
| NASA Earth Science Enterprise: | http://www.earth.nasa.gov/ | ||