Temporally stable patterns in grain yield and soil water on a dryland catena

Grain yields from dryland cropping systems can be affected by water availability, topographic position, and crop rotation. The temporal stability over 14 years of observed spatial patterns is investigated by exploring: (1) relationships between grain yield, soil water content, and topographic position; (2) whether the crop rotation system affects yield; and (3) possible interactions between landscape position and crop rotation. Data have been collected along a soil catena for winter wheat [Triticum aestivum (L.)] and corn [Zea mays (L.)] grain yields and soil water content from 1986 to 1999 as part of a long-term study of sustainable dryland agroecosystem management in eastern Colorado, USA. The experiment included two-year (wheat-fallow), three-year (wheat-corn-fallow), and four-year (wheat-corn-millet [Panicum miliaceum (L.)] or sunflower [Helianthus annus (L.)]-fallow) rotations arranged in a randomized block design (two replicates of each rotation/phase) with repeated measurements from three slope positions. Over all years, analysis of variance revealed a strong effect of topographic position on both corn and wheat yield (p < 0.0001), but no significant effect of crop rotation (p > 0.44) or rotation by position (p > 0.56). Yield at the toeslope was significantly different (p < 0.05) from yields at summit and sideslope positions. The patterns of yield and soil water along the catena were temporally stable, as shown by strong relationships between slope positions. Linear regressions between slope positions explained 69-90% and 65-85% of the spatial-temporal variances in grain yields and profile soil water, respectively. Such spatial relationships can be useful for estimating spatial patterns from either temporal data collection or simulation results at limited spatial locations within a field. (c) 2006 Elsevier Ltd. All rights reserved.