Participants in Working Group Discussions

Facilitator: Bruce Quirk (USGS/EROS Data Center) quirk@edcmail.cr.usgs.gov

Federal perspective: Tom Loveland (USGS/EROS Data Center) loveland@edcmail.cr.usgs.gov

Resource scientist: Jim Merchant (University of Nebraska-Lincoln) jm1000@tan.unl.edu

Mark Anderson (University of Nebraska-Lincoln) manderson@unl.edu

Stephen Boss (University of Arkansas) sboss@comp.uark.edu

John Dixon (University of Arkansas) jcdixon@comp.uark.edu

Rob Dzur (University of Arkansas) rob@cast.uark.edu

Will Gosnold (University of North Dakota) gosnold@badlands.nodak.edu

Ray Hunt (University of Wyoming) erhunt@uwyo.edu

Mark Jakubauskas (University of Oklahoma) jakubaus@ou.edu

Edward Martinko (University of Kansas) e-martinko@ukans.edu

Bruce Maxwell (University of North Dakota) maxwell@cs.und.edu

Stephen Schiller (South Dakota State University schilles@mg.sdstate.edu

John Thomlinson (University of Puerto Rico) thomlins@sunceer.upr.clu.edu

Betty Walter-Shea (University of Nebraska-Lincoln) agme012@unlvm.unl.edu

Timothy Warner (University of West Virginia) twarner2@wvu.edu

Robert Weih (University of Arkansas-Monticello) weih@uamont.edu

Background and Initial Discussion

It is quite evident that human activity has significantly impacted the environment of the Great Plains. The social, economic, political and ecological repercussions of environmental change are increasingly complex and difficult to address. Anthropogenically-induced global warming will, for example, in time, affect biodiversity, water regimes, weather-related disasters, land use and agricultural production in North America. Because Great Plains agroecosystems are especially sensitive to changes in temperature and precipitation patterns, impacts on this region may be especially pronounced. As average temperatures increase, crop regions may shift, drier areas may become more humid, soil erosion may accelerate and so forth. But when and where will such changes occur, what are the risks and potential impacts of change, how do different landscapes respond to short term and long term environmental stresses, how should plans and priorities be established for dealing with the consequences of environmental change?

Such complex problems can only be addressed with complex models and analyses that typically require many types of geospatial data including both environmental (e.g., land cover type, albedo, moisture, canopy architecture, spatial dimensions [patch size, fragmentation], production [biomass] or phenological state), and human factors (e.g., population characteristics, transportation and utility systems, land use). It is also usually critical to consider the temporal dynamics of such factors (e.g., seasonal, interannual or long term change). It is increasingly important that natural resource managers, scientists, public policy-makers and other decision-makers have comprehensive, accurate, current and appropriately-detailed digital geospatial data. They must, as well, have computational tools that will enable them to efficiently and effectively identify, access, archive, move, manipulate, visualize and analyze such data. And they will require technology to support and facilitate collaborative decision-making which may involve interdisciplinary teams working in different locations sharing data over a high speed network.

Land Surface Cover

Information on land surface cover is critical to virtually all Earth System Science research including studies of hydrologic modeling, numerical weather forecasting, global climate change, biodiversity and agriculture. Land surface cover is defined as the material (e.g., vegetation, soil, rock, snow, ice) that is present at a specific geographic location at a specific time of observation. Land cover characterization involves identification and measurement of cover attributes such as albedo, moisture, vegetation physiognomy, leaf area index, canopy architecture, spatial dimensions (patch size, fragmentation), production (biomass) or phenological state. It may also involve assessment of the temporal dynamics of cover (e.g., interannual or seasonal change).

The earth’s land cover, to varying degrees, reflects and integrates environmental factors (e.g., climate, soils, terrain) and anthropogenic activity. Therefore, changes in the character, spatial distribution or temporal dynamics of land cover are often considered indicators of environmental changes. Surface cover is, of course, in a constant state of flux. Short-term changes may stem from seasonal or interannual variability in meteorological conditions; long-term variation may be a harbinger of climate change.

In recent decades it has become practical to monitor surface cover change over large areas using satellite remote sensing systems such as Landsat, SPOT or the NOAA AVHRR. Landsat and SPOT data have relatively high spectral and spatial resolution, and have been found useful for surface cover characterization at local and regional scales. The NOAA AVHRR, on the other hand, provides broad areal coverage at coarse resolution but has high temporal resolution, and has been used to map and monitor land cover at continental and global scales. Land cover is often mapped and assessed using spectral vegetation indices that convey information on vegetation "greenness", a property that has been shown to be broadly indicative of photosynthetic activity and primary productivity, and associated with other ecological processes such as energy and water balances. Both intra-annual and interannual variation in vegetation can readily be observed using satellite imagery.

Although it is possible to map and monitor many surface cover characteristics using remote sensing, our abilities to characterize cover are incomplete. For example, we are currently limited in our abilities to explain and predict how observed large-scale intra-annual and interannual variations in greenness are related to plant phenology, soils, climate and terrain conditions and to anthropogenic activity. If surface cover is to be used as an indicator of environmental change, it is important that we be able to differentiate between short-term anomalies that simply relate to expected seasonal or interannual variability and changes that may reflect long-term "permanent" changes in climate.

Remote sensing and geoinformatics (GIS) are closely linked to one another, and, used in concert, are key tools for mapping and characterizing the land surface. Nevertheless, additional research is needed to establish new and improved methods for characterizing land surface cover using these technologies. Research should be directed towards issues such as calibration and validation of datasets, scaling, multi-source/multiscale data integration, spatial and contextual analysis and methods for exploiting data from new sensors.

Suggested Research Themes and Topics

A Great Plains Research Agenda

The working group believes that the Great Plains should become a key testbed for establishing better understandings of relationships between land surface characterization and remote sensing. The region exhibits substantial diversity, having, for example, a steep temperature gradient north to south and an equally dramatic precipitation gradient east to west across the region. Moreover, the land cover is diverse, and is especially sensitive to interannual meteorological anomalies and disturbances.

Although all topics outlined above are worthy of research, it was agreed that the working group would initially recommend, as its highest priority, establishment of an ESS-based network of long-term observational sites designed to enhance the scientific framework for understanding Great Plains landscapes. It was noted that there is currently a significant disconnect between the relatively abundant remote sensing data available for the region and ground data collection required to support research and analysis. Field data that are currently being collected (e.g., at two LTER sites, agricultural experiment stations) are generally site-specific, not comparable, unique data and are not designed to support remote sensing investigations. In addition, existing field sites do not adequately represent the complete range of Great Plains environments.

The principal objective of this proposed endeavor would be to develop a network of sites spatially distributed over the environmental gradients of the Great Plains to characterize Plains ecosystems in a statistically valid manner. These sites would extend from Mexico to southern Canada and from western Iowa to the Rocky Mountains, including EPSCoR states, other states, and the Canadian Prairies Provinces. The sites would be designed to provide data useful for (1) initialization, calibration, validation and improvement of environmental models and (2) calibrating measurements, improving understanding and explaining observations, and validating analyses obtained via remote sensing. Such data would allow individual research scientists to address many of the themes identified above.

Clearly, establishment and maintenance of such a network is beyond the capabilities of any single institution. Such an effort would have to be a collaborative project involving all states, and many different institutions (including universities, public agencies and non-governmental organizations [NGOs]) in the region. Data would be provided over the Great Plains Network.

The network would not in any way be a replacement for existing long-term field sites such as the Konza Prairie LTER, or the ARM/CART site in southern Kansas and northern Oklahoma. Rather the network would build on and complement existing field data collection efforts. Initially we would conduct an inventory of ongoing field data collection activities to identify established sites. (LTER, ARM, TNC sites, university experiment stations), document instrumentation, data holdings and shortcomings. We would build WWW capabilities to facilitate access to data.

Subsequently we would work to upgrade data collection at existing sites via new instrumentation, new types of data collected, frequency of data collection, better methods for data management, archive and access (via visualization tools) of data and so forth. We would then identify and establish additional sites to capture information on areas of the Great Plains not adequately covered by other sites. Protocols for data collection would be established for all sites to ensure comparability between sites and over time.

Questions That Need to be Addressed

Potential Funding

NSF

NASA

NOAA

EPA

USDA

TNC/Heritage Programs - cooperative ventures

Private foundations

Next Step

Established listserver at ess_landcov@ulysses.unl.edu to continue discussions.