Our research in this area (Section 2B) encompasses spatial patterns of processes in native grassland, responses to manipulations and landuse, and fluxes of trace gases.

In addition to long­term monitoring of net primary production (Table 1.1, Fig. 1.5), we have been very active in issues concerning patterns and controls on primary production, as well as in assessing method-ological problems, developing new techniques, and spanning spatial scale issues (Boindini et al. 1991, Milchunas and Lauenroth 1992, Lauenroth and Sala 1992, Todd et. al 1993, McNaughton et al. in press). Early modeling and analytical work showed that traditional means of estimating NPP could lead to serious errors (Singh et al. 1984, Lauenroth et al. 1986, Sala et al. 1988). A long­term 14C experiment substantiated these findings, and provided realistic estimates of NPP and an alternative to error­prone methods (Milchunas and Lauenroth 1992).

Spatial patterns of aboveground net primary production (ANPP) at the CPER are controlled by location on catenas at a toposequence scale, and soil texture at the larger landscape scale (Liang et al. 1988). Soil texture is also an important control on plant functional types with different rooting distributions (Milchunas et al. 1992, Lee and Lauenroth 1994). Analyses of temporal patterns of ANPP showed a positive relationship with precipitation, but that ANPP was more variable than precipitation (Lauenroth and Sala 1992). An important result from comparing these data to a regional data set was that spatial and temporal relationships between ANPP and precipitation are not interchangeable (Fig. 1.6). Grazing has negative but small effects on ANPP and variability in cool­season precipitation has a large effect on ANPP (Milchunas et al. 1994).

The SGS is characterized by patchy plant cover and associated spatial heterogeneity of soil resources (Hook et al. 1991). Heterogeneity in soil organic matter (SOM) at this scale is as strong as that induced by topography, and is tightly associated with root biomass (Hook et al. 1994, Kelly et al. in press). Enrichment of C and N under plants is associated with perennial bunchgrasses (Vinton and Burke 1995), and develops at a rate that corresponds with plant community successional dynamics (Burke et al. 1995). Death in individual bunchgrasses results in rapid depletion (3 years) of the enriched zone with respect to mineralizable N and labile SOM pools, but heterogeneity of total C and N persists for decades (Fig. 1.8, Kelly and Burke, submitted).

Additional factors in small­scale nutrient distributions involves the influence of aboveground structures on erosion processes, and on the quality of inputs by different plant species. There is substantial accumulation of wind­ and water­transported material under blue grama and plains prickly­pear (Opuntia polyacantha). Grazing increases the movement from interspaces to underplants, and these effects are greatest in summit and midslope topographic positions (Burke et al. unpubl.). Plant species differ in quantity and quality of biomass, but differences in nutrient cycling under their canopies are slight (Vinton 1994). Soils under bunchgrasses had more microbial biomass and activity than under a rhizomatous grass, and soils under annuals had the highest mineralization rates. Comparing the importance of plant cover pattern to that of plant species effects across a gradient from the CPER to Konza, we found that plant cover pattern is most important in SGS, and plant species differences are most important in causing spatial pattern in tallgrass prairie (Vinton 1995).

SOM and nutrient availability are also strongly influenced by both soil texture and topographic position ( Fig. 1.7 , Hook and Burke in prep). In two long­term 15N studies (Delgado et al. 1995 and Hook et al. in prep), we have found that toeslope areas tend to retain up to 90% of 15N added 10years ago, while summits and midslopes retain half or less. Grazed pastures retain less 15N than ungrazed pastures.

Our trace gas flux studies (Mosier et al. 1991, 1993, 1994a,b, 1995, Bronson and Mosier 1994, Parton et al. 1994, Scholes et al. 1994) are confirming that aerobic SGS soils are important sinks for atmospheric CH4, and that gaseous N losses may be important to long­term N balance and productivity. Calculations using IPCC global warming potential indicate that the SGS has a net global warming capacity of ­1100 (sink for greenhouse gases), and N gas loss rates are similar to atmospheric inputs. Fertilization has a large effect on N₂O emissions and CH₄ uptake ( Fig. 1.9).

In our investigation of paleosol geochemistry at the CPER, we proposed a Holocene paleoclimatic scenario (Kelly et al, 1993). A number of factors indicated that the early Holocene was cooler than either the mid­Holocene or current soil forming intervals, and that temperature has not increased from the mid­Holocene to the present (Blecker et al., in press). The 13C/12C ratios of paleosol SOM, phytoliths and CaCO3 established the dominance of C3 vegetation in the soil forming interval spanning 10,000­8,000 y.b.p., and the dominance of C4 vegetation in both the 5,000­3,000 y.b.p. and contemporary soil forming intervals (Kelly et al, 1993). Organic C and phytolith data suggest that both the early­ and mid­Holocene climatic conditions were more favorable for plant productivity than the present climate ( Fig. 1.10 , Kelly et al. submitted). This suggests cooler, moister conditions than presently occur. Pedon development in both the early and mid Holocene paleosols suggests wetter soil moisture regimes than in the present (Blecker, 1993).

Soil water monitoring data emphasize the interannual and seasonal variability in soil water as well as the critical role that soil texture plays in mediating storage and losses Fig. 1.11, Singh et al. in prep). Soil water availability is critical in controlling recruitment of blue grama (and probably other species) by influencing seed production, germination and establishment (Lauenroth et al. 1994, Section 2A). Soil water availability is closely related to ANPP (Singh et al. in prep). Relationships between ANPP and precipitation (see Biogeochemical Processes above) assume that precipitation is a good surrogate for soil water. At large temporal (growing season or larger) and spatial scales (km2 or larger) this is a good assumption.

Water losses (evaporation and transporation) are largely controlled by water availability. Energy to evaporate water is almost always far in excess of the supply as indicated by the annual or seasonal ratio of precipitation to potential evapotranspiration (Parton et al. 1981). Sala et al (1992) found that the probability of a ratio of precipitation to potential evapotranspiration >1 was only 0.1 at the daily scale and decreased exponentially to 0.01 at the monthly scale. Quantifying bare soil evaporation and understanding how it is affected by soil texture are critical issues in understanding water loss. Results from an experiment with mini­lysimeters confirm a 2­stage water loss pattern with a short period of high water loss followed by a long period in which losses decrease exponentially (Wythers et al. in prep). By contrast, the mini­lysimeter results contradict the idea that the depth to which evaporation can remove water from a soil decreases as coarseness of soil texture increases. We found the greatest depth of loss for a coarse sand.

Many of our questions about structure and function of SGS ecosystems require an understanding of long­term patterns of soil water dynamics. We used a simulation model to generate 30 years of soil water data for the CPER to provide this long­term view (Sala et al. 1992). Several key results were identified by this work. First, the general temporal pattern in soil water at the daily scale is not identical to the pattern in precipitation because of an interacting pattern of atmospheric demand. The peak in average soil water precedes the peak in average precipitation by almost a month. Second, the layer that most frequently had available water was very shallow (4­15 cm) ( Fig. 1.12 ). Layers above this were drier because of bare soil evaporation and layers below because most precipitation events are too small to move water that deep in the soil. Third, on average water penetrated to 100 cm (Fig. 1.12 ). In wet years, it reached 135 cm but in dry years it only reached 45 cm. These data provide important information for understanding vegetation structure in the SGS.

Natural disturbances: The majority of natural disturbances are small (<100 m2) and there is an inverse relationship between size and frequency (Coffin and Lauenroth 1989). The most frequent disturbances (cattle fecal pats, harvester ant nest sites, root feeding invertebrates, and burrows from small animals) produce patches that range in size from 5­40 cm diameter ( Fig. 1.13 , Hook et al. 1994) and are important in shaping plant community structure and diversity across the landscape (Coffin and Lauenroth 1990). An important focus of our natural disturbance work has been the effect of disturbances on individuals of blue grama and its subsequent recolonization of disturbed patches. Previous studies had reported that blue grama can not recover after disturbance. In a study designed to look at the effects of patch size and neighbors on blue grama seedling growth and survival, we found that in openings <30 cm diameter, established neighbors could preempt resources and inhibit seedling establishment (Aguilera and Lauenroth 1993). Openings >30 cm were large enough to result in seedling establishment. Hook et al. (1994) reached a similar conclusion about the size of a regeneration gap based upon how the root systems of neighbors explored patches created by disturbances ( Fig. 1.14 ). Simulation analyses and assessments of recovery on abandoned fields (see below) confirm that blue grama does reestablish following a disturbance although in some situations the rate may be quite slow ( Fig. 1.15 , Coffin and Lauenroth 1990, Coffin et al. 1996, Lauenroth et al. 1994). Recovery rates are dependent upon the characteristics of the disturbance, and in particular size and soil texture (Coffin and Lauenroth 1994, Coffin et al. 1993).

Root feeding by white grubs (June Beetle larvae) can cause mortality of blue grama over areas ranging from 0.1 to 100s of m2. A long­term experiment to follow recovery on grazed and ungrazed areas affected by white grubs (Coffin et al. submitted, Table 1.1, Section 2E) has produced the following results: 1. successional dynamics of plant functional types on patches affected by white grubs were similar to other disturbances; 2. strong relationships between survival of blue grama on the affected patches and cover on ungrazed areas indicate the importance of initial conditions; 3. the importance of grazing increased over time and the importance of initial conditions decreased.

Human­induced disturbances:
We are continuing our research on the recovery of plants and soils on abandoned agricultural fields along precipitation and temperature gradients in Pawnee National Grasslands and soil texture gradients at the CPER (Burke et al. 1995, Coffin et al. 1996). We found a large variability in recovery of the vegetation that could not be explained by climatic factors. Recovery may be related to soil texture, since the ability of blue grama to recover through seedling establishment is affected by silt content of the soil (Lauenroth et al. 1994). Net N mineralization and other indicators of active SOM were lowest on cultivated fields, but were not significantly different between abandoned and native fields. Recovery of SOM in abandoned fields appears to involve accumulation of C and N under perennial plants. Higher N mineralization and turnover in cultivated fields may make them more susceptible to N losses (Ihori et al. 1995a), and recovery of N cycling in abandoned fields appears to involve a return to slower N turnover and tighter N cycling similar to native SGS. Although variation in native soil C and N correlated with climate and soil texture (Ihori et al. 1995b), soil losses due to cultivation were not explained by these variables. Rates of recovery were small compared with loss rates due to cultivation( Fig 1.16 ).