• I. Introduction
  • A. Conceptual Framework
  • B. Shortgrass Steppe Site Definition
• II. Continuing and Proposed Research
  • A. Populations and Trophic Dynamics
  • B. Biogeochemical Processes
  • C. Paleoecology/Paleopedology
  • D. Water and Energy Exchange
  • E. Natural and Human Induced Disturbance
  • F. Synthesis
  • G. Summary

In this Section we describe our conceptual framework, long- and short-term experiments, and how they fit into our conceptual framework. We have utilized tables and figures to summarize this information. Table 1.1 summarizes our long-term experiments, including their design and sampling frequency. Published experiments are described in less detail, with references listed. We use figures and tables to highlight progress from past and ongoing research; numbered figures are cited in the text, un-numbered figures provide important additional information. Finally, our complete progress report from the past 6 years of LTER funding is on our World Wide Web page.

Our conceptual framework asserts that one must consider the interplay of several forces, which occur at a variety of spatial and temporal scales, in order to understand the structure and function of SGS ecosystems. There are five components that we have identified as particularly important in shaping the SGS: climate, natural disturbance, physiography, human use, and biotic interactions ( Fig. 2.2 ). Below, we provide an overview of the SGS in order to frame the unique interactions of these components, and then elaborate on each in turn.

The SGS occupies the middle of the productivity gradient along which the LTER grassland sites lie. It is unique among North American grasslands for its long evolutionary history of intense selection by both drought and herbivory, leading to an ecosystem that is very well adapted to withstand grazing by domestic livestock (Mack and Thompson 1982).

The distinctive features of the SGS are both its vegetation and the concentration of biological activity and organic matter belowground. The vegetation of the SGS is characterized by low- growing plants that are either tolerant or resistant to grazing and drought. The most common plants are two species of C4 grasses, a cactus, and several species of dwarf-shrubs, all of which have a large fraction of their biomass belowground. The vegetation is strongly dominated by Bouteloua gracilis (blue grama), the grass species with the greatest tolerance to grazing and drought (Weaver and Albertson 1944). This species contributes 60-80% of the plant cover, biomass, and net primary productivity of shortgrass ecosystems. The large concentration of biological activity belowground reflects the distribution of plant production (Milchunas and Lauenroth 1992) and the enhanced rates of energy flow through heterotrophs belowground (Lauenroth and Milchunas 1992). It is also explained in part by the fact that most biologically active elements in grasslands are protected from natural disturbances by being stored in soil organic matter (SOM).

This distribution of biotic components, with a large bias towards the belowground portion of the system, leads to important predictions about the response of SGS ecosystems to disturbances: those events that largely affect aboveground plant biomass, such as fire or grazing, are not likely to cause changes in total organic matter, or large alterations in the biotic community ( Fig. 2.5 ). Recovery from such disturbances may occur in as few as 2-3 years due to the tolerance of the biota and the high rate of turnover ofaboveground plant biomass. In other types of grasslands, there may be shifts in plant species composition in response to grazing or drought, leading to alterations in biota, and ecosystem structure and function. However, the evolutionary adaptations of organisms in the SGS provide additional resistance to grazing.

Our research indicates that the biota have a large response to interannual and seasonal variations in precipitation (Lauenroth and Sala 1992). Perennial C₄ bunchgrasses maintain their presence and cover in all years, changing slowly with dry and wet periods. By contrast, C₃ grasses and annual forbs respond significantly to individual wet or dry years. Abundance and species composition of small mammals reflect precipitation- induced changes in vegetation structure (Grant et al. 1977) and resulting fluctuations in production of seeds and arthropods (Crawford 1991). Ecosystem functions such as net primary productivity (Lauenroth and Sala 1992), net N mineralization (Hook and Burke in prep), and trace gas flux (Mosier et al. 1991) are very responsive to interannual and seasonal variation in precipitation.

Temperature plays a lesser role in influencing SGS ecosystems because of its relatively small range of variability compared to precipitation. We have recently initiated a long-term experiment to evaluate the influence of increased temperatures predicted by climate change scenarios on ecosystem structure and function (see Section 2.II.B ). There are significant feedbacks from SGS ecosystems to atmospheric processes, including reflected radiation, water vapor, and trace gases (see Section 2.II.D).

In distinguishing our 5 components, we exclude both drought and grazing from our list of natural disturbances due to the long evolutionary history of these selection pressures in the SGS. The climate of the SGS has fluctuated among subhumid, semiarid, and arid over the past 10,000 years (Table 2.1). At the time of settlement, the SGS was the home of native Americans and large numbers of bison, prairie dogs, elk, pronghorn, bighorn sheep, and deer. These herbivores had been present since the retreat of the Pleistocene glaciers, 10,000 years ago. Conceptually, we list herbivory except by cattle under biotic interactions, and cattle grazing under human use; drought is dealt with under atmospheric processes.

Soil texture and depth play an important role in controlling water availability, a key constraint for the biota of SGS ecosystems. In sandy soils, water moves to greater depths and surface evaporation is lower than in fine-textured soils, leading to lower bare-soil evaporation rates and different vegetation structure (Noy-Meir 1973, Lane 1995). Deep sandy soils often support large shrubs which, in turn, strongly influence the presence and abundance of small mammals (Stapp in prep). Soil texture also exerts control over decomposition and N mineralization ( Fig. 1.4). Areas of fine texture are characterized by higher SOM accumulation, due to the lower mass-specific decomposition rates, and leading to higher N mineralization (Hook and Burke in prep) and N2O flux (Mosier et al. submitted).

The SGS has been subjected to several periods of conversion to croplands and subsequent periods of cropland abandonment. Sixty percent of the SGS is currently being cultivated (Lauenroth et al. 1994) and up to 30% more was cultivated at some time in the past 100 years. As a result, a component of our studies deals with the recovery of abandoned croplands, although our research focuses on native steppe. Cropping drastically alters biological diversity (e.g. Holland and Coleman 1984), reduces SOM (Haas et al. 1957, Burke et al. 1989), and alters the temporal and spatial distribution of plant biomass. Our work on abandoned fields indicates that while vegetation, nematodes, active SOM, CH₄ consumption and N₂O emissions may recover within 50 years, total SOM does not.

In summary, our conceptual model ( Fig. 2.2 ) suggests that the current state and vulnerability of SGS ecosystems (structure and function) requires understanding of five major components and their interactions with the ecosystem: climate, natural disturbance, physiography, human use, and biotic interactions. Our past, present and future research is focused on these components even though we have not organized this proposal explicitly around them.

The SGS LTER project has been based at the Central Plains Experimental Range (CPER), a 6500-ha research site established by the USDA in 1937. The site was also a U.S. IBP site in the late 1960's and early 1970's. The work by USDA scientists and the IBP experience and database have been tremendous resources for the SGS LTER project (Table 1.4).

A recent analysis (Burke and Lauenroth 1993) suggested that the precipitation, temperature, vegetation, landuse, and soil characteristics of the CPER represent a small portion of the SGS. The areas not adequately represented include fine-textured soils and areas with mean annual precipitation >321 mm. To increase our realm of inference to better represent the SGS, and to increase the land area that we may use in our studies, we have increased the spatial extent of our LTER site to include both the CPER and the Pawnee National Grasslands (PNG) (Table 2.2). The PNG represents 78,100 ha of public lands administered by the U.S. Forest Service (USFS) adjacent to the CPER, and extending 90 km to the east. The PNG is discontinuously distributed across northeastern CO because these lands are the result of acquisitions of private land beginning in the Dust Bowl era. By expanding to our newly defined site, we increase our realm of inference to 23% of the SGS ( Fig. 2.7 ; see also Fig. 5.2).

We have an agreement with the PNG based upon the Memorandum of Understanding between the USFS and NSF, and upon a letter of agreement that we have exchanged (Appendix 1). This new partnership has the potential to dramatically improve our ability to meet both our LTER mission and that of the USFS with respect to understanding and managing the SGS.

Grasses, shrubs, forbs, and succulents are the major components of the vegetation structure of the SGS. One grass species, blue grama, contributes a huge fraction of the biomass and NPP. Because of its overwhelming importance, our past plant population work has focused heavily on blue grama. This work is closely related to our disturbance research ( Section 2.II.E ). Another grass species, Buchloe dactyloides (buffalo grass) can achieve co-dominant status on certain locations; we have recently begun to study its demography and interactions with blue grama. We are proposing to continue to study blue grama and buffalo grass as well as to initiate studies on the most important forb and shrub species. We are proposing to initiate long-term studies of population dynamics of Opuntia polyacantha (plains prickly-pear). This species makes up a relatively small proportion of biomass or NPP but because of its growth form and protection from grazing, it has the potential to have a large effect on ecosystem dynamics under cattle grazing.

Recruitment of blue grama is an infrequent event controlled by the availability of viable seed, temperature, soil water, and competition from established plants (Lauenroth et al. 1994, Aguilera and Lauenroth 1995). Requirements for germination and establishment are well known; we therefore focus our efforts on other processes. To date we have learned that (1) seed production is negatively related to soil clay content, and the shape of the relationship is different for grazed and ungrazed locations (Coffin and Lauenroth 1992); (2) seed storage in the soil is very limited and variable in time and space (Coffin and Lauenroth 1989); and (3) the suitability of a site for germination and establishment depends strongly on proximity to an established individual of blue grama (Aguilera and Lauenroth 1993, 1995, Hook et al. 1994). We are also currently exploring the roles of wind and cattle on seed dispersal.

We have also conducted a simulation analysis using a soil water dynamics model (SOILWAT) to evaluate the frequency of climatic conditions suitable for the establishment of the blue grama seedlings. We found that the probability of establishment increases as soil silt content increases and that return times for establishment events range from <30 years (silty clay loam) to >5000 years (sand) (Lauenroth et al. 1994). These results lead us to design new experiments to test the hypothesis that seedling establishment of blue grama is related to soil texture, and in particular silt content. We will collect soils of different textures, place them in pots in a growth chamber, and water them with different amounts and frequencies of addition. Seeds collected from the CPER will be added to each pot. Germination and establishment will be determined for each soil texture and water treatment. At the end of the study, seedlings will be harvested for above- and belowground biomass as well as number and length of roots. We will also conduct a long-term field study adding blue grama seeds to a variety of microsite types (in tussocks and in openings of several sizes) at locations with different soil textures. Seeds will be added to each site each year and the number of surviving seedlings will be counted at regular intervals.

Little is known about mortality in long-lived clonal plants such as blue grama. Our field studies on the role of small disturbances suggested that mortality occurs only if the entire plant is affected by the disturbance ( Fig. 2.8 ). To further investigate factors causing mortality of blue grama and the role of environmental variability on recruitment and growth, we will initiate a new study using permanent plots (0.5 m2) to evaluate individual and population dynamics through time. Five plots will be established on each of 2-4 soil textures in moderately grazed pastures and exclosures. Individual plants of all species will be mapped in each plot annually. Plots will be selected to include blue grama as well as other species. This study will also provide information about the demography of the co-dominant buffalo grass.

Genotypes of blue grama that differ with respect to culm height, leaf length, culm number, and other key characters will be identified and collected in the fall from the LTER common garden. They will be cloned over winter in a greenhouse to produce the required number of plants. In the spring they will be transplanted into four cattle grazing treatments: grazed, protected, protected but clipped, and protected by prickly-pear. Measurements will be made on growth and fitness. If genotypes exhibiting alternative morphologies show changes in fitness rank across grazing environments, then a genotype x environment interaction model of the maintenance of genetic variation in blue grama will be supported.

We will initiate a study of long-term population dynamics of prickly-pear using 20 similar sized clones (> 30 cm dia.) in each of 3 moderately grazed pastures and exclosures. Annually, a drawing of each clone will be updated with information about the numbers of live (new and surviving) and dead cladodes (pads). We will also record which cladodes support a flower. This will be critical information for understanding the dynamics of clones as well as the role that prickly-pear plays in shortgrass ecosystems.

We will begin a short-term study to investigate the influence of prickly-pear on plant species richness, seed production and soil seed storage under moderate grazing and across several soil textures. We will sample 20 similar sized clones (> 30 cm dia.) in 3 moderately grazed pastures. Within each clone and in an adjacent space between clones, we will identify all plant species, harvest inflorescences, and take soil cores from which we will germinate stored seeds in a greenhouse.

We will address the hypothesis that grazing by both large (cattle) and small (prairie dogs) herbivores will result in on-town plant communities with a high diversity of native plant species relative to off-town communities and low numbers of exotic invaders compared to plant communities grazed only by prairie dogs. This will entail protecting plots on prairie dog towns with cages with different mesh sizes; large mesh, open-sided cages allowing prairie dog access and excluding cattle, and small mesh cages that exclude both prairie dogs and cattle. The initial design would include 4 treatments: 1) off-town/cattle grazed, 2) on-town, grazed by both, 3) on-town, no cattle, and possibly 4) on-town, both excluded.

Prairie dogs are potential prey species for mammalian (coyotes, red or swift foxes, and badgers) and avian (Swainson's hawks, great-horned owls, prairie falcons) predators, representing a patchy, concentrated, and predictable (within-season) prey source. We plan to investigate the relative frequency of prairie dogs in diets of these predators. We expect that the relative frequency will be high for predators near colonies, and will decline with distance to colonies. The scale of decline will vary with predator type.

To determine which predators utilize prairie dogs as a significant dietary source, we will review datasets on the food habits of local predators. We have identified an unpublished dataset that includes fecal and pellet material of mammalian and avian predators that was collected by a variety of investigators but as yet has not been analyzed. We will analyze these remains for the presence of prairie dog fur and/or bones using standard scat and pellet analysis techniques (Reynolds and Aebischer 1991, and Zimmerman et al. In press). The proportion of scats and pellets containing prairie dog remains will be used in conjunction with a Monte Carlo simulation to estimate the sample sizes necessary to detect, at a specified level of reliability, a real difference in each species' diet from scat and pellets collected at three different distances (<0.5 km, 0.5-1.0 km and > 1.0 km) from the prairie dog town (Reynolds and Aebisher 1991).

Two procedures will be used to collect the appropriate sample sizes of feces and pellets within these distance categories. During the growing season when feces and pellets are difficult to detect in the landscape, they will be collected at regular intervals from raptor nest sites and carnivore den sites. The distance from the closest prairie dog town will be used to categorize the remains. During the winter when fecal and pellet detectability is high, we will walk 2-3 km line transects that radiate from prairie dog towns with random orientations. All scat and pellets will be collected and identified to predator species based on scat and pellet morphometrics (Green and Flinders 1981, Danner and Dodd 1982) and comparisons with material collected at den and nest sites. Sampling intensity during both seasons will be based on the results of the power analyses and fecal and pellet densities. All scats and pellets will be analyzed for the presence of prairie dog remains. The effect of distance from a town on the proportion of prairie dogs in the diet of each predator species will be analyzed using logistic regression (Hosmer and Lemeshow 1989).

In 1994, we began to annually census populations of nocturnal lagomorphs. We spotlight rabbits along a 32-km route of gravel roads and trails on one night in Jan, Apr, Jul, and Oct. Perpendicular distances from the vehicle to rabbits are recorded; these data are used to estimate density. Black- tailed jackrabbits (Lepus californicus) and desert cottontails (Sylvilagus audubonii) are the most abundant species ( Fig. 2.15 ). White-tailed jackrabbits (L. townsendii) are relatively uncommon and occur primarily on upland prairie.

Warming, wetting, and N additions LT14
A prediction of our conceptual framework is that there are two-way interactions between the structure and function of SGS ecosystems. Results from a long-term data set on ANPP suggested that ecosystem structure constrained ANPP (Lauenroth and Sala 1992). Production was greater in dry years and less in wet years than expected by comparison with sites with mean annual precipitation corresponding to the wet and dry conditions. How does the relationship between ecosystem structure and function change under altered climate and resource availability?

Seasonal soil-atmosphere exchange of CH4, N2O and CO2 LT15
Since 1990, we have been sampling trace gas flux (CH₄, N₂O and CO₂) at 10 locations weekly on the CPER throughout the year (Table 1.1). Most trace-gas budgets assume that trace-gas exchange between the soil and the atmosphere stops when soil is snow covered or soil temperatures drop to 0oC. Since soil temperatures are at this temperature or below for a large part of each year, we have chosen to continue measurements throughout the winter. Our results to date indicate that winter fluxes of NO, N₂O, and CH₄ uptake make up 10-50% of the annual mean flux ( Fig. 1.9).

Plant Production and Nutrient Cycling Across Soil Textural Gradients NEW-LT
Several experiments were set up during the last 6 years to evaluate the impact of soil texture and nutrient cycling on plant production. The results from the comparison of sandy loam with sandy soil supported theinverse texture hypothesis (Noy Meir 1973) with higher aboveground production in the sandy soils. However, our comparison of plant production along catenas from sandy loam uplands to loamy swales showed that production was higher in the swale sites ( Fig. 1.5 ). The most likely causes of higher production in the swales are 1) increased soil water availability and 2) increased N availability. Data suggest that N mineralization in swales is higher because of greater N retention in fine textured soils (Delgado et al. 1995) and deposition of SOM. The only explanation for higher water availability in swales would be downslope movement, but our neutron probe soil water data are not resolved enough to detect this movement. We plan to seek additional funding to conduct an experiment to evaluate the relative importance of nutrient availability and water availability on the observed spatial patterns of plant production.

Long-term 14C plots, belowground production, and root biomass dynamics LT16, LT17, NEW-LT
We plan to continue to evaluate belowground production using a radioisotope technique. Large plots (8 m2) were labeled with 14C to assess the implications of short- and long-term carbon dynamics on estimates of aboveground, crown, and root production using 14C dilution, 14C turnover, and traditional harvest methods (Fig. 2.17) (Milchunas and Lauenroth 1992). 14C turnover appears to provide reliable estimates of aboveground, crown, and root production, although they are an average over several years. In 1995 we collected our 11th annual samples from the 14C plots.

NEW:
Equipment to sample minirhizotron tubes was obtained through a Colorado Agricultural Experiment Station equipment grant in 1995. We began this past growing season installing tubes adjacent to the 14C plots and the root-harvest plots, where we have 11 yrs of root biomass data collected through the growing season. Minirhizotron tubes will also be installed at our 6-site, 4- treatment grazing experiment, and in 2 shrub-dominated areas.

Short-term study on C distribution and decomposition throughout the soil profile NEW-ST
A recent comparison of root biomass data (Fig. 2.18)(Lee and Lauenroth 1994) and SOM data (Yonker et al. 1988) suggested that soil carbon profiles do not match those of root production or root biomass; the proportion of SOM at depth is significantly higher than one would expect from grass-dominated ecosystems. We will utilize the 14C plots to determine where the 14C is going in terms of labile and recalcitrant soil C pools. Results will allow us to infer the rate of SOM stabilization throughout the soil profile. In addition, we will conduct a litterbag study to evaluate decomposition rates with depth. We will incubate root litterbags at 7 depths in 10 replicate profiles, to be collected at intervals of 4 mo., 1, 2, 3, 5, 10, and 20 years. This work will be guided by hypothesis 2.6 (Table 2.4).

Long term 15N studies in the SGS LT18, LT19
In 1988, we initiated a long-term 15N study to evaluate the interaction of topographic position and grazing on N retention and distribution ( Fig. 2.19 ). We labeled 3 1-m2 plots on summits and toeslopes in each of 3 catenas, and in two grazing treatments, heavily grazed and protected. We plan to continue to sample these treatments on a decadal scale for an indefinite period of time. We sample each plot to determine 15N retained in plant material (leaves, crowns, and roots), and in mineralizable, slow (particulate SOM, Cambardella and Elliott 1992), and recalcitrant fractions of SOM. This study will be important to assess the interactions of abiotic variables and management in controlling N retention. Our work on N retention is guided by hypothesis 2.7 (Table 2.4).

The influence of plant functional type on N retention NEW-ST
We are currently testing the influence of dominance by C₃ and C₄ plants on N retention. We have located two sites on the CPER that are co-dominated by C₃ and C₄ plants, and have initiated two experiments on these sites. Our hypothesis is that C₃ dominated plots will have highest losses of N in summer, when the rate of plant uptake is lowest, C₄ dominated plots will have highest rates of N losses during spring before plants are active, and co-dominated plots will sustain the lowest rates of N loss. We established 3 replicate plots of community type at each site, for each of 3 treatments: control, spring water addition, and summer water addition. Each plot has been augmented with 2 g15N m-2, and is being sampled at the end of the 1995 and 1996 growing seasons to assess N retention. Trace gas flux (N₂O, NO, and CH₄) is measured in each treatment on a weekly basis, from April through October ( Fig. 2.20).

Study of enhanced CO₂ on SGS ecosystem structure and function LT20
Atmospheric CO₂ concentrations have been rising over the past several decades at unprecedented rates, and are projected to continue rising (IPCC 1995). No field studies have addressed how elevated CO₂ might impact the SGS. A study such as this will be particularly interesting because the SGS is water-limited and C-enrichment may increase productivity, and because the SGS is very near the ecotone of dominance by C₃ and C₄ plant functional types which respond differently to enhanced CO₂. Members of our PI's recently received funding to initiate a field and simulation experiment to examine the influence of enhanced CO₂. We are coupling the use of open-top chambers for field CO₂ enrichment, soil N cycling and trace gas flux measurements with measurements on the physiological responses of two plant species, blue grama (C₄) and western wheatgrass (Agropyron smithii, C₃). We will incorporate new knowledge gained from the experiments into simulation models to extrapolate our results in time and space across the grasslands.

Interactions of plants, topography, and grazing on SOM and N availability LT21
During the past 6 years, we put a great deal of energy into evaluating the influence of individual plants on C and N cycling. Hypothesis 2.9 has guided our work in this area (Table 2.4). We conducted numerous studies that indicate that the patchy distribution of individual plants results in high spatial heterogeneity of soil resources ( Fig. 2.13 )(Hook et al. 1991, 1994, Burke et al. 1995, Vinton and Burke 1995, Vinton 1994), comparable or even higher than that at the landscape scale. Our results indicate that spatial heterogeneity is large, but is associated with SOM pools that change relatively rapidly, leading to both rapid development and decay of this small-scale patterning. We recently initiated a long-term study to evaluate the influence of individual plants on SOM and N availability. We are characterizing the interactions of plant species, topographic location, and grazing intensity on spatial heterogeneity of soil resources. We are sampling 3 landscape positions (summit, midslope, and toeslope), 3 micropositions (under blue grama, under prickly-pear, and between plants), and in two grazing treatments (heavily grazed and ungrazed), with 5 samples in each combination of treatments. In each location, we are measuring microtopography to estimate erosion/deposition, and sampling surface soil to assess the relative roles of organic and mineral deposition. We plan to sample at decadal intervals to assess the actual rates of accumulation and loss of material ( Fig. 2.21).

The role of individual plants and community composition during succession (see Section E)

Atmospheric inputs NEW-LT
One of the most important indicators of ecosystem changes is alteration of element inputs. Although our site has long been a member of the National Atmospheric Deposition Program, there are still several key areas of element balance that we do not understand. First, what is the relative contribution of weathering vs. atmospheric inputs? Second, what is total N deposition, including both wet and dry deposition? Our work on this topic will be guided by hypothesis 2.10 (Table 2.4)

Dry deposition is an important component of total atmospheric deposition (Lovett 1994). Current estimates of N deposition based upon precipitation underestimate the contribution of atmospheric N to the SGS. Reliable estimates of dry deposition will be useful in constructing more accurate N budgets and provide baseline data on the possible effects of increasing urban development on N deposition in the SGS. We plan to develop a dry N deposition program to provide baseline information and begin long-term sampling of atmospheric N inputs. The mass flux of N to the plant/soil surface will be estimated using a modified Bowen-ratio method with measured changes in temperature and nitric acid concentration with altitude (Huebert and Robert 1985).

We have used the stable C isotope composition of paleosol SOM, phytoliths and CaCO₃ to estimate the relative proportions of C₃ and C₄ plants growing during soil forming intervals and made inferences to prevailing climate. We measured the quantity of SOM and phytoliths in paleosols and compared these data to contemporary soils to indicate relative climatic differences. Our results indicate that C₃ plants were dominant during the soil forming interval 10000-8000 YBP and that C₄ plants were dominant during both the 5000-3500 YBP and contemporary soil forming intervals (Kelly et al. 1993) ( Fig. 1.10, Table 2.1). This suggests the early Holocene was cooler than mid Holocene or present. Organic C and phytolith mass is greater in early and mid Holocene paleosols than in contemporary soils, which suggests a wetter climate than present (Blecker et al. in press). Hypothesis 2.12 (Table 2.5) guides this work.

Evaluation of phytoliths from the mid Holocene soil forming intervals NEW-ST
Although there is significant overlap in phytolith form among grass genera, three subfamilies can be reliably identified: Chloridoid, Festucoid and Panicoid (Twiss et al, 1969; Fredlund et al, 1985). The Chloridoid subfamily represents three tribes, Chlorideae, Erogrosteae and Sporoboleae. The Festucoid subfamily represents four tribes, Festuceae, Hordeae, Aveneae, and Agrostideae. The Panicoid subfamily represents three tribes, Andropogoneae, Paniceae and Maydeae. Each of these groups corresponds to a unique set of environmental (temperature and moisture) conditions.

Phytoliths will be extracted from the silt and fine sand sized fractions of 14C dated paleosols. Phytoliths in both contemporary vegetation and soils will be characterized morphologically as reference samples. Existing phytolith classifications will also be used to identify grass taxa (Twiss et al, 1969; Brown, 1984). The stable C isotope composition of phytoliths and SOM from 14C dated paleosols will be quantified to determine the relative importance of C₃ and C₄ plants present at the time the phytolith was deposited.

CPER paleosols appear to have been preserved as small "patches" (<100m2 - 1 km2) rather than as extensive, continuous surfaces. It is unclear, given their contemporary landscape positions, what positions they occupied on the paleolandscape (Hypothesis 2.13, Table 2.5). A robust interpretation of the Holocene paleoenvironment requires the determination of paleosol area and morphometry to determine if paleosols represent relic fluvial features with distinctive microclimates. A subset of representative paleosols from each of the soil forming intervals will be cored to determine their area and morphometry.

Madole (1994) stated that, under the present climate, eastern Colorado is near the threshold of widespread mobilization of sediment by wind. We will estimate the area of paleosols for the PNG by correlating paleosol occurrence with stratigraphy and physiography. We will compare paleosol maps with existing data on location of sand sheets and other eolian features as well as soil survey information to assess the potential for paleosol exhumation (Hypothesis 2.14, Table 2.5).

We have used stable C isotope composition of SOM, phytoliths and CaCO₃ to estimate the distribution of C₃ and C₄ plants during Holocene soil-forming intervals. This information, in turn, has been used to make inference to Holocene climate. Our regional interpretations of climate, to date, are based on data of limited geographic extent. We will date and characterize paleosols across the PNG to develop a more robust interpretation of regional climate change. Further, we will utilize published stratigraphic and archaeologic data pertaining to this area to refine our interpretations (Hypothesis 2.15, Table 2.5).

Meteorological monitoring and water balance LT22, LT23
Microclimate: We have an automated microclimatic station that has been functional since 1982. It is at a site on which microclimatic monitoring was begun in 1970 under the US-IBP program. In addition to the automated system, we make manual observations that also link to a data set begun in 1970. Our automated system makes hourly measurements of precipitation, wind direction, wind speed at 3 heights (1, 2, and 4 m), air temperature, surface temperature (infrared), dew point temperature, vapor pressure, total radiation, soil water at 3 depths (fiberglass blocks; 5, 10, and 20 cm) soil heat flux, and soil temperature at 4 depths (2.5, 5, 10, 20 cm). Manually we record air temperature, relative humidity, precipitation, open pan evaporation, and soil temperature at 7 depths (1" 2.25" 4" 8" 20" 40" 72"). Observations are made once daily in the morning. We plan to continue both automated and observations.