PE&RS October 2000

Building on Three Decades of Remote Sensing and Decision Support: The NASA Great Plains RESAC and the Kansas Applied Remote Sensing (KARS) Program

Introduction

The rapid growth of information technologies has provided exciting new sources of data, tools, models, and other information to environmental managers and planners at all levels in the public and private sectors. A major challenge facing the scientific community in the 21st century is to incorporate remote sensing technology as key components of the decision-making and planning processes. To accomplish this goal, decision support systems must increasingly utilize spatial information from the myriad of remote sensing sources to provide meaningful answers to the questions confronting decision-makers.

In 1998, the Kansas Applied Remote Sensing (KARS) Program, at the University of Kansas (Lawrence), was designated as one of nine NASA Regional Earth Science Applications Centers (RESACs).

The mission of the RESACs is to apply remote sensing and related technologies to issues and problems of regional significance. This article highlights the key contributions of the Great Plains RESAC and the KARS Program to remote sensing, outlines our major research initiatives, and summarizes some of the many decision support and management products, services, and applications we are developing.

Forging Partnerships For Decision Support

KARS was established in 1972 by NASA and the State of Kansas to conduct research on techniques that enable public agencies and private firms to better utilize data from satellite remote sensing technologies and related geographic information systems (GIS) in decision-making, policy formulation, and planning. To date, the KARS Program has completed more than 150 projects that have involved land use/land cover inventory, rangeland and forest characterization, wildlife habitat evaluation, mapping of irrigated lands, surface mined lands inventory, soil conservation needs assessment, wetlands inventory, and urban area analysis.

The KARS Program has expertise in both landscape-level and close-range remote sensing. The program has 16 research and professional staff members in addition to 10 graduate students. KARS staff includes specialists with backgrounds in geography, ecology, biology, computer science, mathematics, agriculture, environmental studies, and natural resources management which reflects the inter-disciplinary approach used to apply remote sensing and GIS technologies. KARS staff also have extensive experience in providing both national and international workshops, short courses, and training sessions on remote sensing and GIS techniques.

The foundation of the Great Plains RESAC is the KARS Program, whose activities in remote sensing and decision support span more than three decades. The current direction and progress of the Great Plains RESAC and the KARS Program in bridging the divide between the academic, and private and public sectors rests squarely upon the shoulders of that longevity.

The NASA Great Plains RESAC

RESACs were established by NASA to utilize public/private consortia as end-to-end partnerships among the research, service and end-user communities to address region-specific studies and user-defined problems. Consortium participants are involved at each stage of the process of problem resolution: from identifying user needs, research, data analysis and reporting, to decision-making. The Great Plains RESAC Consortium includes members from seven private companies, five Federal agencies, four state agencies, three local government agencies, three universities, and three regional partnerships.

Region-specific studies conducted by the RESACs address economic and policy issues in the region identified by the user community through the Regional Workshops of the U.S. Global Change Research Program. For the Great Plains, the agricultural economy is the linchpin of the region and a key component of the U.S. economy. The Great Plains RESAC therefore focuses on three areas important to the economy of the Great Plains and the nation: assessing grassland condition and productivity, monitoring and projecting crop production and yield, and monitoring land use/land cover change.

User-defined problems addressed by the RESACs are identified by a specific user community located in a broad region (e.g., Great Plains) and include an applications problem or set of related problems facing the user community and decision-makers in the region. Because of the central role of agriculture in the Great Plains, the Great Plains RESAC focuses on the following user-defined issues: vegetation damage assessment and agricultural production modeling, land use and land cover mapping, land use change analysis, agricultural land evaluation, global climate change analysis, and habitat modeling. Decision Support for State and Local Government Statewide Landcover Mapping. KARS has provided three land cover maps of the state of Kansas. Each of these comprehensive, high-quality maps was developed in direct response to user-identified needs as a basis for decision-making. To date the State of Kansas has invested more than $2 million in development and utilization of land cover information derived from remote sensing data. In 1974, the Kansas Department of Economic Development funded creation of a land cover map of Kansas. The “Kansas Land-Use Patterns: Summer 1973” map was visually interpreted from 18 Landsat MSS images. The resulting 1:1,000,000-scale map depicted twelve land use and land cover classes based on a modified Anderson Level II classification scheme. Despite the large map units, the use of mixed classes (e.g., rangeland with non-irrigated crop), the age of the data, and the fact that it was never converted to a digital form, requests for this map from a variety of end users continue to the present. An update of this map was undertaken in 1991 when the Kansas Land Cover Mapping Project was funded by the Kansas Water Plan under the auspices of the State GIS Policy Board. Completed in 1993, the map was digitally classified from single-date Landsat TM imagery and depicts general USGS Anderson Level I land use and land cover classes with a 2-acre minimum map unit. The imagery from this work is featured in a “Satellite View of Kansas” (Fig 1). The digital database used to create the map is available for download through DASC at http://gisdasc.kgs.ukans.edu. A third comprehensive statewide map of Kansas is in the final stage of preparation (Fig 2). The map has been compiled in cooperation with the USGS Gap Analysis Program. GAP is a nationwide conservation and management program, implemented state-by-state, that uses land cover maps as a basis for habitat modeling and species protection.

The KARS Program has a long history of pro-active leadership and participation in cooperative partnerships with the public and private sector. This record of accomplishment includes several milestones: Public Sector Kansas Interagency Task Force on Applied Remote Sensing. Organized by KARS in early 1981, the Kansas Interagency Task Force on Applied Remote Sensing was composed of 11 agencies representing state, county, and local government. A primary objective of the Task Force was to enhance interagency communication, coordination, and cooperation on statewide use of remote sensing and GIS technologies. Kansas Commission on Applied Remote Sensing. The success of the Task Force led the Kansas Legislature in 1984 to enact legislation establishing a permanent Kansas Commission on Applied Remote Sensing (KCARS). KCARS accomplished several important tasks, including the establishment of a State Map Coordinating Committee and initiation of a study for the development of a Kansas Geographic Information Center. Kansas GIS Policy Board. KCARS was in effect the predecessor of the current, highly successful Kansas GIS Policy Board. The Board’s Data Access and Support Center (DASC) is the incarnation of KCARS’ vision for a Kansas geographic information center. The Board is chaired by the Kansas Water Office and has 22 members, including representatives from local, state and Federal government agencies and private industry. With an annual database development budget of $250,000 from the Kansas Water Plan, the Board and its subcommittees establish long-range objectives and plans, formulate and enact data standards, determine database development needs, and operate the DASC. KARS, through its parent organization, the Kansas Biological Survey, is an active member of the Kansas GIS Policy Board and several of its subcommittees. The KARS Program has been designated by the Board to develop, maintain, and update the digital land use/land cover database for the State of Kansas. Private Sector The GreenReport®. The GreenReport® was the first commercial remote sensing product developed by the KARS Program. The GreenReport® is an AVHRR-based weekly product that provides timely information about crop progress and vegetation condition within the conterminous United States. Originally released in 1995, the GreenReport® has recently been enhanced through the work of the Great Plains RESAC as a result of user feedback from the agribusiness community. TerraMetrics, Inc., AquaSolu-tions, LC., and Planalytics, Inc. The product development and technology transfer activities of the Great Plains RESAC and the KARS Program have spawned the development of two new Kansas-based companies, TerraMetrics, Inc. (TMI), and AquaSolutions, LC. (ASLC). RESAC activities have also spurred the creation of a new joint venture company, Planalytics, Inc. (Wayne, PA). These private sector companies are developing programs and services to market and distribute the remote sensing/geographic information systems products developed by the Great Plains RESAC. The KARS Program has developed several commercial products for use in agriculture, fisheries and forestry, that are currently being distributed and marketed through TMI, ASLC, and Planalytics, Inc.

Development of the Kansas GAP land cover map has received funding from a broad range of Federal and state partner agencies. The land cover phase of Gap Analysis for Kansas, which began in 1995, uses multi-date Landsat TM imagery to map grassland, woodland, and wetland areas at the vegetation alliance level. This mapping effort uses a hybrid unsupervised/supervised classification procedure that relies on field data collected from more than 3500 sites.

In preparation for future updates of the digital land cover database and maps for the state of Kansas, KARS has begun work in cooperation with the Kansas Biological Survey, Department of Wildlife and Parks, and the State GIS Policy Board to create an ongoing multi-season, multi-year image data set.

The archive will provide up-to-date imagery for future land cover maps. The image data set will also provide a valuable resource for the Kansas Department of Revenue in determining land valuation, for the Kansas Water Office in managing water resources, and for the Kansas Department of Wildlife and Parks to augment the more than 20 cooperative remote sensing projects completed to date.

CRP Program Support Tools. Another avenue of inquiry for land cover mapping at KARS has focused on the U.S. Conservation Reserve Program (CRP). The CRP program has resulted in conversion of 14.8 million ha (36.5 million acres) of cropland to grassland, woodland, or other conservation uses. KARS researchers have used multi-season imagery to develop highly accurate maps of cropland and grassland for the pre- and post-CRP enrollment periods. These maps were used to identify regions of cropland that had been converted to CRP grassland between years. The CRP maps have also been used to evaluate soil characteristics of CRP grasslands and the effect of CRP lands on landscape structure and wildlife habitat.

Groundwater Management Support Tools. One of the most pressing issues facing agricultural producers on the Great Plains is the use of groundwater for irrigation of crops. In order to create decision-support tools for groundwater managers and agricultural producers, we are investigating the use of multi-date satellite imagery to accurately distinguish irrigated from non-irrigated crops. This represents a challenging problem in those areas of the Great Plains where irrigated and non-irrigated crops of the same type (e.g., wheat, corn, grain sorghum) are grown side-by-side. Initial results using multi-date imagery classification are of sufficient accuracy to permit modeling the distribution of irrigated cropland by individual crop type.

Fig 3. The GreenReport® mapset. Green indicates increased vegetation; gold shows decreased vegetation; brown means no change. A) NDVI map showing biweekly-composited greenness values for July 7-20, 2000. B) Comparison of mid-July 2000 to early-July 2000. Evident is the green-up of the Corn Belt and dry conditions along the Rocky Mt. Region. C) Comparison of mid-July 2000 to mid-July 1999. D) Mid-July 2000 compared to the ten-year average for mid-July.

Geospatial Solutions for Agriculture The task of taking fundamental research and transferring it out of the hands of satellite remote sensing specialists and into those of real world farmers and agricultural resource managers is a formidable one, and the challenges are numerous. KARS, however, has achieved an important level of success in this endeavor. Since the late 1980s, KARS has developed several new remote sensing applications for large scale agricultural monitoring and crop modeling. As a NASA RESAC, we are forging partnerships that will allow the transfer of these applications into the hands of end users. Fundamental research is the foundation of applications development: In the late 1980s, KARS researchers were looking at multi-date measurements of the spectral characteristics of various land cover and management types. It was noted that the spectral response patterns for different kinds of grasslands varied across the growing season. This initial observation led to a series of NASA funded projects in the 1990s that resulted in the development of multitemporal analysis methods for Landsat Thematic Mapper (TM) and Advanced Very High Resolution Radiometer (AVHRR) imagery. In particular, we began developing applications for the AVHRR Normalized Difference Vegetation Index (NDVI) Composite data set produced by EROS Data Center. During the past five years, KARS scientists have also fostered the development of harmonic (Fourier) analysis and related time-series techniques using multi-temporal resolution imagery. We have used harmonic analysis of time-series AVHRR data to identify crop and grassland types for the western US Great Plains such as corn, winter wheat, alfalfa, and native prairie grasslands. Translating fundamental research into real world applications. Central questions that confront the agricultural community are: How do current crop conditions compare to conditions observed the previous week? How do current crop conditions compare to crop conditions last year at this time? How do current conditions compare to the average condition of past crops during the same time period?

To address these questions, we began analyzing AVHRR biweekly data sets to develop statistical methods for comparing AVHRR data.

The set of “relative difference” maps that were developed eventually became known as the GreenReport® (Fig 3), now commercially available through TerraMetrics, Inc. and Planalytics. Thousands of subscribers receive these weekly map products. A free version of the GreenReport® is available at www.kars.ukans.edu.

Crop yield monitoring and forecasting tools and applications. We are now developing crop yield forecasting models for corn, winter wheat, soybean, and several other crops for use throughout the conterminous U.S. This work builds upon the idea that multitemporal NDVI profiles can be used to assess plant development status both within and across growing seasons.

KARS crop yield forecasting models have evolved into highly accurate tools that:

• Provide county level crop yield estimates months ahead of harvest
• Provide accurate yield estimates well in advance of the USDA crop yield estimates

Similar models will be developed for monitoring agricultural productivity globally. We are working with our consortium partners and the private sector to develop products that will address this need. We are also beginning to investigate the relationships between long-range weather trends and remotely sensed data as a basis for longer range yield forecasting.

Long-range forecasting is important to agribusiness because lead-time is particularly critical for business planning. Long-range weather forecasting will also allow the preparation of crop timeline forecasts (e.g., planting, harvest, peak season dates) and related strategic planning tools for seed dealers, fertilizer dealers, large scale grain purchasers and related agribusiness activities.

Forest Management Applications

In May 1999, we were selected by NASA to develop methods that use remote-sensing data and advanced geostatistical methods to create maps of forest age and successional state, as well as key forest characteristics, such as density, biomass, leaf area, basal area, and height.

During the past decade, we have worked on multiple projects involving lodgepole pine communities in the Greater Yellowstone Area (GYA). Research has included characterization and modeling of lodgepole pine seral stages, analysis of image texture and forest regeneration, effects of insect damage on spectral response in a forest stand, investigation of forest structure / age class in Yellowstone and Grand Teton National Parks, and generating forest canopy structure models from remotely sensed data (Fig 4).

Using a geostatistical approach to image analysis, we are developing two demonstration projects that will show the use of remote sensing for insect damage assessment and mapping forest cover types. This approach will provide significant advantages in managing forest resources, especially forests that are in remote or inaccessible locations.

We are implementing several methods to assure the transfer of this technology from research and development to the user community. First, geostatistical software modules that interface with current remote sensing analysis systems are being developed. Second, a web site containing online tutorials for geostatistics and remote sensing, as well as online demonstrations of the analysis procedures can be accessed at www.kars.ukans.edu/forest/. Furthermore, the software modules and sample field and image data sets used in this project will be available for download.

This integrated remote sensing/geostatistics approach will reduce the amount of time and the costs required for forest inventory and mapping. The resulting maps will provide information on forest characteristics that were previously very difficult or impossible to map (e.g. leaf area index (LAI) and aboveground biomass).

By monitoring changes in forest characteristics over time, forest managers can use this technique to map insect defoliation, wildfire damage, and forest regeneration.

Global Initiatives

KARS researchers have worked with resource managers and government officials around the world to conduct research and develop decision support tools for a diversity of management issues. Supported by funding from the National Science Foundation and the World Bank, KARS has recent or ongoing work in locales ranging from China and Mongolia, El Salvador and Mexico, to Kazakhstan and Zambia. Not only has our work resulted in scientific advances, but it has also provided key opportunities for showing other governments how remote sensing can aid decision-making and allocation of resources.

In Central Asia and Mongolia, for example, we recently worked with Colorado State University to develop applications for remotely sensed data that assess the impacts of climatic change on environmental stability. As a testimony to the power of technology transfer and outreach, the Chinese government recently awarded our Chinese research partners funding to continue this work.

In Zambia, we have been working with the country’s government and academic officials for the past six years to establish Zambia’s first National Center for Remote Sensing. When the livestock sector in Kazakhstan suffered economic collapse, KARS used a remote sensing approach to identify the land use change and climate factors that were responsible. In Mexico, we are examining the potential impacts of climate change on the biodiversity of small mammals and birds throughout the country. In El Salvador, we are developing remote sensing applications for assessing the impacts of coffee growing practices on forest biodiversity.

We continue to expand our international relations and research activities by hosting visiting scholars from China, Africa, Mexico, and Central America. In addition, government officials from countries throughout the world have visited KARS to learn more about how remotely sensed data can be used to develop decision support tools that will address their unique natural and agricultural resource management challenges.

A Look Ahead

The Great Plains RESAC and the KARS Program will continue to build on more than three decades of remote sensing research and technology transfer to the public and private sectors. We will continue to refine and develop new vegetation phenology metrics for estimating grassland and cropland productivity and new methodologies and decision-support tools for land cover mapping and analysis.

As they become available, we will utilize new sensors to provide unique information to support our research and product development activities in response to the needs of agricultural and natural resources decision-makers.

The challenge to provide meaningful spatial information derived from remote sensing technologies as a key component of the decision-making process represents a major goal for the 21st century. This goal can only be accomplished by delivering meaningful information in response to well-defined problems of end user communities through effective decision support systems. The Great Plains RESAC and the KARS Program look forward to meeting this challenge.

Selected References

Egbert, S.L., Yang, R.Y., Price, K.P., Nellis, M.D., and R. Boyce. 1998. Mapping Conservation Reserve Program (CRP) Lands Using Multi-Seasonal Thematic Mapper Imagery. Geocarto International 13(4):17-24.

Lee, ReYang, D.L. Kastens, K.P. Price, R. Lee, and E.A. Martinko. 2000. Forecasting Corn Yield in Iowa Using Remotely Sensed Data and Vegetation Phenology Information. Proceedings, of the Second international conference on Geospatial Information in Agriculture and Forestry, Lake Buena Vista, Florida, January 10-12.

Jakubauskas, M.E. and K.P. Price. 1997. Empirical relationships between biotic and spectral factors of Yellowstone lodgepole pine forests. Photogrammetric Engineering and Remote Sensing 63(12):1375-138.

Price, K.P., and M.E Jakubauskas. 1998. Spectral retrogression and insect damage in lodgepole pine forests. International Journal of Remote Sensing 19(8): 1627-1632.

Stewart, A.M., S.L. Egbert, C.L. Lauver, E.A. Martinko, K.P. Price, D.L. Peterson, S. Park, C.F. Blodgett, and J.F. Cully. 2000. Land cover mapping for Gap: a hybrid classification approach to mapping the vegetation of Kansas. Proceedings of the Annual ASPRS Meetings, Washington, D.C., May 24-26.

Whistler, J.L., S.L. Egbert, E.A. Martinko, D.W. Baumgartner, R. Lee, and M.E. Jakubauskas. 1996. “Development of the Kansas Digital Land Use/Land Cover Map From Satellite Multispectral Imagery.” In Morain, S. and Baros, L. (eds.) Raster Imagery in Geographic Information Systems, pp. 328-334.

For more information:

Theresa J. Crooks
Great Plains RESAC Project
Coordinator
KARS Program
Phone: 785-864-7369
E-mail: darwinsfinch@crooks.net

Top

Home