The semi-arid regions of the Pacific Northwest are characterized by a high degree of annual temperature and precipitation variation. As a result of this climatic variation, dryland nitrogen fertilizer trials on fallow- ,wheat rotations typically demonstrate a variable response. Wheat growers in the area must not only cope with this climatic variation and its sundry effects upon their livelihood, they must also make decisions regarding the future level of anticipated climatic variation. The specific objectives were to: 1) develop a climatically responsive yield potential prediction model for soft white winter wheat from historical data at the Sherman Branch Experiment Station (Moro, OR); 2) modify this model for use on commercial fields; 3) field simulate five fallow-crop precipitation patterns characteristic of the variation found in the Sherman county area of eastern Oregon in order to test the yield potential model: 4) examine the effects of precipitation variation on nitrogen fertilizer responses, moisture storage and depletion and nitrogen mineralization; and 5) establish a quantitative relationship between precipitation/ soil moisture and nitrate accumulation in both the fallow and crop seasons. Two interacting regression models were developed to estimate grain yield levels in the 250-350 mm precipitation zone of eastern Oregon. The first model estimates yield potential from monthly precipitation and temperature values. The second model estimates the percent grain reduction due to delayed crop emergence. The grain yield model was adapted to commercial fields using a Productivity Index factor (PI). The PI is a measure of the productivity of other locations in relation to the Sherman Branch Experiment Station, using water-use-efficiency (WUE) as the basis of comparison. The field simulation of five fallow-crop precipitation patterns demonstrated that the maximum grain yield response occurred at 40 kg N (soil + fertilizer)/metric ton. The grain yield model demonstrated a 15% level of accuracy on a commercial field basis in both field trials and a survey of past production levels (1972-1980). It was hypothesized that the distribution of precipitation in the fallow and crop periods had an effect on both the amount and distribution of stored soil moisture. The field simulation demonstrated that more soil moisture was stored at the 90-240 cm depths by the patterns with more fallow season precipitation when measured in March of the crop year. Soil moisture storage and storage efficiencies fluctuated throughout the fallow and crop periods. At the cessation of the winter precipitation season in both the fallow and crop periods (March), the storage efficiency was highest when low levels of precipitation occurred. At this point in time, the mean crop period storage efficiency was 10% below the mean fallow period storage efficiency (34 and 44%, respectively) in both simulation studies. Soil moisture, temperature and immobilization requirements of crop residues interact to affect the net amount of nitrogen mineralization. The mineralization model proposed by Stanford and Smith (1972) was tested under field conditions. When the nitrogen immobilization requirement of the crop residues was included, the actual and predicted values were in agreement at the close of the 1978 fallow period. A nitrogen deficit was predicted at the 0-30 cm depth at the close of the 1980 fallow; however, the actual levels indicated a net accumulation of nitrate-nitrogen. Crop season mineralization, inferred from Mitscherlick and a-value extrapolations, in 1979 demonstrated that there was a decreasing amount of net mineralization during the crop season with increasing amounts of both fallow and crop season precipitation. Crop season mineralization in 1980 indicated that there was no net accumulation of nitrogen, rather a tie-up of 14 kg N/ha. This result reflects both the unsatisfied immobilization requirement predicted for the 1979 fallow season and crop season denitrification. / Graduation date: 1981