Linking plant growth responses across topographic gradients in tallgrass prairie

TitleLinking plant growth responses across topographic gradients in tallgrass prairie
Publication TypeJournal Article
Year of Publication2011
AuthorsNippert, JB, Ocheltree, TW, Skibbe, AM, Kangas, LC, Ham, JM, Shonkwiler-Arnold, KB, Brunsell, N
Pagination1131 -1142
Accession NumberKNZ001387
KeywordsANPP, Eddy covariance, Flux footprint, LAI, Mesic grassland, topography

Aboveground biomass in grasslands varies according to landscape gradients in resource availability and seasonal patterns of growth. Using a transect spanning a topographic gradient in annually burned ungrazed tallgrass prairie, we measured changes in the height of four abundant C4 grass species, LAI, biomass, and cumulative carbon flux using two closely located eddy flux towers. We hypothesized that seasonal patterns of plant growth would be similar across the gradient, but the magnitude of growth and biomass accumulation would vary by topographic position, reflecting spatial differences in microclimate, slope, elevation, and soil depth. Thus, identifying and measuring local growth responses according to topographic variability should significantly improve landscape predictions of aboveground biomass. For most of the growth variables measured, classifying topography into four positions best captured the inherent spatial variability. Biomass produced, seasonal LAI and species height increased from the upland and break positions to the slope and lowland. Similarly, cumulative carbon flux in 2008 was greater in lowland versus upland tower locations (difference of 64 g m−2 by DOY 272). Differences in growth by topographic position reflected increased production of flowering culms by Andropogon gerardii and Sorghastrum nutans in lowland. Varying growth responses by these species may be a significant driver of biomass and carbon flux differences by topographic position, at least for wet years. Using a digital elevation model to classify the watershed into topographic positions, we performed a geographically weighted regression to predict landscape biomass. The minimum and maximum predictions of aboveground biomass for this watershed had a large range (86–393 t per 40.4 ha), illustrating the drastic spatial variability in growth within this annually-burned grassland.