Characterizing the temporal stability of crop yield across heterogeneous field conditions is an important goal of precision agriculture. In the dryland cropping region of Eastern Washington and Northern Idaho, terrain and soil attributes are important determinants of crop yield and similar field-scale patterns in relative performance are often observed. Our objective was to quantify annual variations in crop yield to assess the stability of spatial and temporal patterns. A non-aligned, systematic sampling grid consisting of 369 geo-referenced points over 37 ha was established in 1999 at the Washington State University Cunningham Agronomy Farm near Pullman, WA. Various crops (wheat, barley, canola and peas) were grown from 1999-2003 and yield was hand harvested (2-sq. m) at each geo-referenced point every year. Spatial and temporal analyses consisted of geostatistical techniques and conditional simulation. These analyses demonstrated the marked spatial and temporal stability of crop yield and aided the definition and characterization of crop performance zones across diverse field conditions.