Factors underlying grain yield spatiotemporal variability in California rice fields

The dissertation consists of three chapters. In the first chapter, I determined the spatial and temporal structure of grain yield from 4 years of yield data in two rice fields. Three different regularly spaced sampling grids were tested for their capacity to capture yield spatial variability using variography and the relative information criterion. The grid density necessary to capture the variability changed according to the yield spatial characteristics of each year, making it hard to predict from any one year's yield map which would be the most appropriate grid resolution to use next year. Temporal variability was determined by two different approaches: (1) computing the variance among years of the original, trend and residual yield values from fixed points in the field; and (2) by using cluster analysis of the standardized trend yield values. Cluster analysis reduced the considerable complexity in a sequence of yields maps of these fields to a few general patterns of among year's variations with a given spatial distribution.;In Chapter two, I evaluated several approaches to determine the factors associated with the spatial and/or temporal variability described in Chapter one. These approaches include (a) classical inferential statistics like Pearson's correlation coefficients and stepwise multiple linear regression, (b) non parametric statistics like CART, Mantel and partial Mantel tests and (c) geostatistical analysis like variograms and cross-variograms. I found that each method provided different information, and only by integrating classical inferential statistics, non parametric statistics and geostatistical analyses it was possible to conform a better understanding of causes of yield spatial variability.;In Chapter three I assess the spatial variability in water temperature in two rice fields, quantified water temperature (T_w) effects on yield; and determined whether water temperature can be based on thermal infrared images. I showed that the deleterious effects of cold water on rice productivity occur well beyond the visually impacted area immediately adjacent to the water intake. Spatial variability in water temperature was discernible using remotely sensed thermal infrared images, which are well correlated to ground based measurements.