Landscape Patterns in Soil‐Plant Water Relations and Primary Production in Tallgrass Prairie

Landscape variation in soil water relations, leaf xylem pressure potential (°) and leaf—level net photosynthesis (A) in Andropogon gerardii, and net primary production (NPP) were evaluated during the 1989 and 1990 growing seasons in a northeast Kansas (USA) tallgrass prairie. Landscape patterns were assessed along transects that spanned upland and lowland topographic positions in an annually burned and a long—term unburned watershed. Landscape variability in volumetric soil water content (°) was significantly greater in the unburned watershed (coefficient of variation [CV] = 0.425 and 0.479 for 0—15 and 0—30 cm soil depths in unburned prairie vs. 0.285 and 0.330 for similar depths in the burned watershed). In both watersheds, significantly higher ° and total soil water content (0—30 cm) were measured in lowlands compared to uplands. Topographic anomalies, such as a lowland ridge, resulted in local, small—scale variation in soil moisture that equaled watershed variation. Variation across landscapes in predawn °, which was expected to reflect soil water content, was similar in both watersheds (CV = 0.312). Variation in midday @j was significantly greater across the burned than the unburned watershed in 1990 (maximum range in @j from uplands to lowlands was 0.708 MPa at predawn and 0.662 MPa at midday in the burned watershed). In both watersheds, variation in midday @j was much lower relative to °. Landscape patterns in leaf—level A in A. gerardii, the dominant species in this tallgrass prairie, were inconsistent when upland and lowland sites were compared. During an extended period of drought, A was significantly higher in plants in the unburned watershed. In both watersheds, NPP was strongly correlated with °. However, variability in NPP across topographic gradients in the unburned watershed was much less pronounced (CV = 0.224—0.245) than in the annually burned watershed (CV = 0.364—0.430). Moreover, the slope of the relationship between NPP and ° was significantly greater in the annually burned watershed. We propose that relatively uniform energy limitations across topographic gradients in unburned tallgass prairie, caused by detrital accumulation that absorbs/reflects sunlight, reduced topographic variability in NPP in unburned watersheds. This pattern occurred in unburned watersheds despite greater landscape variation in ° relative to burned watersheds. Analysis of long—term records of NPP from several watersheds supported the hypothesis that variability in NPP associated with topographic position is lower in unburned vs. burned watersheds. Variability in @j across watersheds and between years was muted by negative feedback of canopy leaf area (transpiring surface) on plant—soil water relations. We concluded that patterns in landscape variability in A and @j which may vary significantly over short time intervals, were not good predictors of seasonal carbon dioxide exchange or productivity in this tallgrass prairie. Nonetheless, interactions between A and @j, when combined with nitrogen and energy limitations to A. provide the mechanisms for integrated responses measured across these landscapes.