Avocado yields are decreased by chloride toxicity and soil salinity throughout California orchards but there is little information on the extent to which different rootstocks can be used to improve tree performance under saline conditions. This research has been aimed at the development of a production function model that can be used to predict the impacts of irrigation water chloride content and salinity (EC) on avocado yields. The model further evaluates the effects of different soil chemical and physical properties, water chemistry, and rootstocks on the accumulation of chloride in the leaf tissue. Data have been collected from 10 orchards that span the major avocado production areas from San Diego to San Luis Obispo. Our modeling approach involves the use of an artificial neural network (ANN) program that enables us to separate out complex interactions that cannot be detected using traditional statistical procedures. The long term goal of this research is to develop a smart program that growers can access via the internet to predict how different water qualities and soil properties will affect their yields. This will also provide guidance on the best rootstocks for different irrigation water qualities, and estimates of yield losses and gains under different management scenarios.
