We have collected data for numerous cross-site studies this year. We are using a 15N tracer to assess the importance of soil organic matter (SOM) and soil texture on N retention in native grasslands (Barrett and burke 1998). In the spring of 1996, we applied 50 and 2.5g of 15N labeled NH4SO4-N to 1 m² plots on coarse and fine textured soils at 5 sites along a temperature gradient between the panhandle of Texas and southeastern Montana. We collected soils and plant material at the end of the 1996 and 1997 growing seasons to estimate N retention of plant biomass and SOM pools, and have just completed sample 15N analysis. Retention of the low addition of N was close to 100% for all sites. Fine textured soils and soils with high SOM content retained more 15N than coarse textured soils and sites with low SOM content. Our results suggest that soils can be an important sink of anthropogenic nitrogen in grassland ecosystems.

In a second study, we have been utilizing several semiarid sites to evaluate how plant life form influences nutrient cycling (Gill and Burke 1997, 1998). Many semiarid ecosystems have experienced a change in dominant pant life form. In the SW United States, many former grasslands and savannas have had an increase in woody vegetations in the last century, and many shrublands in the Great Basin of the United States have been invaded by annual grasses. Associated with these changes are shifts in the dominant rooting patterns and aboveground plant structure. We utilized soil isotopic techniques to assess how changes in plant life form alter the depth distribution of soil organic matter (SOM) at sites in southern Texas, New Mexico, and Utah. We found that in all three invasions, woody vegetation had higher SOM content in surface soils than did grasses, and that annual grasses influence soils to only 30 cm, but there is evidence of perennial grass influence to a depth of 1 meter.

Third, in conjunction with an LTER cross-site grant, small mammal exclosures (lagomorphs and rodents) were constructed in 1997 at three of six SGS large-herbivore exclosure sites. This will allow us to compare the roles of small selective herbivore to large generalist herbivores in the shortgrass steppe, and across sites that differ in productivity and evolutionary history of grazing. Other sites included in the study are Konza, Cedar Creek, Deseret montane, Deseret bunchgrass, and Badlands. Baseline data for ANPP, plant species cover, and soil C&N were collected in 1996. In 1997, baseline data was also collected for shrub density and volume by measuring every individual shrub in the small mammal exclosures and in equal areas of the new large mammal exclosures and adjacent long-term grazed areas. Previous casual observations suggest that small mammals may have large impacts on shrub abundance and structure in this system through winter browsing activities. Cover and ANPP estimates after the first year of treatment will be collected early August 1998.

Fourth, in collaboration with Imanuel Noy-Meir and the U. S.-Israel BSF, a comparative study of the importance of natural grazing refugia in U. S. and Israel grasslands was initiated in 1998. The grasslands in Israel and shortgrass steppe of North America both have a long history of grazing by domestic animals, but geologically formed grazing refugia are common in Israel and rare in the U.S. Great Plains. The role of these refugia in harboring rare plant species was examined. A second aspect involves the potential of Opuntia polyacantha cacti to provide short-term refugia from grazing. The potential interaction between large-scale refugia with a large seed pool and the small-scale cacti refugia is being examined by Dr. Salvador Rebollo of Alcala University, Spain. Preliminary analyses of the data indicate that cacti refugia do not influence plant species richness, but have a large influence on seed production. Detailed analyses of these data are currently underway.

Fifth, although funding for our cross-site project on decomposition and productivity with our Argentinian collaborators has ceased, we are continuing our filed studies and manipulations of water, increased temperature, and nitrogen additions. We recently submitted our progress report for this study.

Finally, we have recently received funding as an LTER supplement to conduct new transect studies from the SGS-LTER to the Konza LTER site, and to initiate new cross-site work with the Sevilleta beginning with a workshop at the Sevilleta this fall. The work with the Konza will follow up on work by Vinton and Burke (1997) that suggested that there are significant changes in plant-soil interactions across the precipitation gradient.

We continue to conduct many regional analyses as part of our larger grassland region research group. SGS-LTER investigators were recently funded by EPA to continue our regional analysis; the LTER and EPA activities are very closely related with respect to our research questions, but the EPA project extends our questions to address the cropland systems that are such an important part of the region. Under LTER funding, we have recently completed two regional analyses of plant species distributions and productivity (Epstein et al. 1997, 1998), and one on the similarity of grasslands and shrublands in N. and S. America (Paruelo et al. 1998). We recently completed an analysis of landuse patterns, and found that environmental variables explain 81% of the variance in landuse management in the Central Grasslands region (Burke et al. 1998). A study comparing primary productivity of native grasslands and wheat fields across regional gradients was completed (Lauenroth et al. submitted). Finally, a graduate student recently conducted a cross-site analysis of IBP and LTER data to assess the patterns in litter quality across large scale climatic gradients (Murphy et al. submitted).

The coupling of the climate version of the Regional Atmospheric modeling System (RAMS) and the biogeochemical model (CENTURY) has been successfully completed (Liston and Pielke in prep). Annual runs with the coupled system have been performed and show different results than when the models are run separately from each other (i.e. with the vegetation condition specified for RAMS, and, alternatively, the weather conditions specified for CENTURY). In addition, a new biophysics module (LEAF-2) has been introduced into RAMS which permits more realistic partitioning of surface heat and moisture fluxes from spatial scales which are smaller than the meteorological model grid.

3. Synthesis

Modeling continues to be an important synthesis activity for the SGS-LTER investigators. We have developed and tested a new daily version of the CENTURY model (DAYCENT) (Parton et al. submitted). The monthly version of the CENTURY nutrient cycling, organic matter cycling, and plant growth submodels were modified to run using a daily time step and the trace gas model (NGAS) was incorporated into the nutrient cycling submodel. We have written a series of papers describing the structure of the DAYCENT model and comparing the model results with observed data. The comparisons included observed vs. simulated soil temperature and water, plant production, nutrient mineralization rates, soil respiration rates and trace gas fluxes.

Among our highlights for synthesis, we synthesized data concerning bird and rodent community and species responses to livestock grazing from studies across North America (Milchuna et al. in press). Dissimilarities of bird community species composition in response to differences in grazing intensities were generally higher in systems with a short than a long evolutionary history of grazing. Preliminary assessment of the rodent data indicates a strong and generally negative effect of grazing on richness and diversity of rodent species in most plant communities.