1a.Objectives (from AD-416)
Develop an intra-seasonal model of grape production that evaluates the economic impacts of different leaching and deficit irrigation strategies on grape production within a single season and across seasons. Analyze the econnomic impact of reduced water suplies and increased salinity on field and farm-level management decisions.
1b.Approach (from AD-416)
The research will use the results of a series of field studies on cooperating farms, laboratory studies at UC Davis, greenhouse studies and sand tank studies at U.S. Salinity Laboratory in the model development.
This agreement supports objective 1 of the parent project. Progress was made on the development of an intra-seasonal model of grape production to evaluate the economic impacts of different leaching and deficit irrigation strategies on grape production within a single season, and development of an inter-seasonal model of grape production that evaluates the economic impacts of different leaching and deficit irrigation strategies across seasons. The mathematical model of grape production represents production as a function of water applications, crop age, and soil salinity in a way that keeps track of the effects of field management on yields and profits over time. Crop history is captured by a permanent biomass index which varies temporally due to biological factors and management practices. The behavior of the biomass variable is calibrated currently to the available evidence from viticultural science literature in arid and semi-arid environments. The model uses a detailed intra-seasonal model of hydrological and soil processes to generate data that relates seasonal field management over the season to the carryover effects on grape vine health and potential impact on future yields. The intra-seasonal model currently uses Dynamic Programming (DP) methods to choose the optimal water use over the life cycle of the crop given the ways in which age, water history, and soil salinity affect crop yields and quality. The DP problem consists of a system with 3 state variables and 1 control variable over a 40-year planning period. Because DP problems are solved recursively, a system with multiple state variables requires optimizing over a large number of possible solutions which can lead to long run times or even a failure to find a solution. A code was written in Mathematica to distribute computations across parallel processors and use pre-compiled functions where possible to decrease program run time. The lead scientist confers regularly with the investigator through email, conference calls and annual meetings.