Evaluating the Potential for Precision Management in Irrigated Rice Fields.

Our current understanding of the mechanisms driving spatiotemporal variability of rice yield is insufficient for accurate prediction and effective management at the sub-field scale.  The spatiotemporal variability of yield, flood water solute concentrations and temperature, and soil properties within four rice fields (i.e. F1, F2, F3, F4) representative of varying soil types and locations within the primary rice growing region in California were quantified and characterized. Mean grain yield and coefficient of variation (CV) at the grid point locations within each field ranged from 9.2 to 12.1 Mg ha^-1 and from 7.1 to 14.5%, respectively. Spatial patterns of soil properties were generally temporally stable. Most notably, soil phosphorous (P) in F3 ranged from 3.2 to 18.1 mg kg^-1 (CV = 49.4%), and soil electrical conductivity (EC) in F4, ranged from 0.6 to 4.8 dS m^-1 (CV = 70.1%). The trend analysis showed that soil nutrients (i.e. dissolved organic carbon, total nitrogen and potassium) were moved across the fields in irrigation water and deposited in the lowermost basins or in areas with low lateral flow. The effect of cold water temperature and land-leveling on yield variability was not observed as others have found. Using a k-means clustering and randomization method, temporally stable low-yielding zones were identified in F1, F3 and F4, whereas F2 lacked a consistent yield pattern. Using Classification and Regression Trees to determine the underlying causes of the yield clusters, the low-yielding clusters were related to EC greater than 0.87 dS m^-1 in F1, EC greater than 2.0 dS m^-1 and P less than 5.95 mg kg^-1 in F3, and soil P below 4.30 mg kg^-1 in F4. Considering the stability of yield and the yield-limiting factors in F1, F3 and F4, each field could be divided into two zones of below-average and above-average yield, and managed independently.