| The drought monitoring products currently available are too coarse, both spatially and temporally, for most agricultural applications. These shortcomings were addressed by developing a near-real-time Agricultural Drought Monitoring System (ADMS) for Delaware.; A statistical yield model and a physiological (CERES-Maize) yield model were calibrated and compared using field trial data from Kent County, DE to determine which is the most appropriate for predicting corn yield. Although the statistical yield model was highly correlated with the observed yields during the calibration period, it did not perform well during the validation period. The results indicated that statistical models are not appropriate for simulating yield at the field-level because they are not able to simulate yield during years with extreme weather conditions and they are location specific.; The sensitivity of the CERES-Maize model was quantified for variations in row spacing, planting density, planting date, initial soil moisture, carbon dioxide concentration, air temperature, solar radiation, and soil type. The results indicate that as row spacing increases and planting density decreases, yields generally decrease. Increasing the carbon dioxide concentration and solar radiation both resulted in greater yields. Initial soil moisture (on January 1) does not have a significant impact on yield. Both air temperature and planting date have a non-linear impact on yield since they only affect yield during years when precipitation is unequally distributed throughout the growing season. Yield is most sensitive to variations in soil type. In some years, variations in soil type produced yield variations of nearly 300%. The amount of precipitation and the timing of precipitation in relation to the crop growth stages determined the magnitude of the soil-type influence. Soil type did not affect yield when precipitation was abundant and evenly distributed over the growing season. However, soil type had a major impact on yield during dry years or when precipitation was distributed such that it was lacking during the moisture sensitive growth stages.; The soil moisture component of the CERES-Maize model was validated with observed soil moisture data. The model provides a reasonable approximation of soil moisture in the upper layers of the soil after May 1. However, there were significant differences in model performance between 2002 (a drier year) and 2004 (a wetter year) and the model has difficulties simulating soil moisture in the lowest layer of the soil and during the first few months of the year. (Abstract shortened by UMI.)... |
