OBJECTIVES

Our objective is to create a decision aid to support recommended soil nutrient management practices for forest trees, especially plantation trees, in Hawaii. The computer software system would integrate into the University of Hawaii's ADSC Fertility Advice and Consulting System (FACS) and be used by agents and clients determining fertilization and other nutrient management systems for trees. As with the FAC system, users would enter soil type and plant (tree) species. For forest trees, growth stage of the tree and estimated growth rate would also be entered. The proposed decision aid would then recommend a fertilization regime, including amount of mineral nutrient required and type of fertilizer. Recommendations would be based on soil nutrient supplying power, tree uptake, and economic efficiency.

APPROACH

In creating the decision aid, we will need to model nutrient demand by forest trees, nutrient supply by Hawaiian soils, and predictions of growth and yield so as to make economic decisions. Nutrient demand will be primarily based on the predicted growth rate of the trees and whether the trees are nitrogen-fixing or not. Nutrient supplying power of the Hawaiian soils is already well-known and encapsulated in the FAC system. Yield of the trees will be estimated from previous tests in Hawaii or results from literature if Hawaii data is unavailable. Fertilizer recommendations are based on the nutrient supplying capacity of the soil and the plant demand. Extensive data for the nutrients available from many of Hawaii's soils are already encoded in the FACS. We plan to use the FAC system and database to model the nutrient supplying power of Hawaiian soils. The FAC system 2 does not, however, take into account the nutrient demand of forest trees. We propose using data from experiments in Hawaii and elsewhere in the tropics to develop nutrient demand functions for forest trees. We will model tree nutrient demand in four phases: nursery, establishment, rapid growth, and maturity. In each phase, we will base nutrient demand on predicted tree growth, less nutrients recycled from litterfall and root dieback. For the proposed project, we will concentrate on the establishment and rapid growth phases. Estimates of potential tree growth will be taken from the literature. We will initially rely on growth rates of trees and nitrogen-fixation to estimate nutrient demand. Further refinements to the model may include nutrient recycling through litterfall, re-translocation of nutrients within the tree, and efficiency of nutrient uptake, if information is available for tree species. Our goal, however, is to estimate nutrient demand for tree species for which published data are rare or non-existent. The forest fertilization module will be added to FACS in order for forest recommendations to be developed and disseminated as part of the existing sample submission, analysis, and reporting stream current in place for annual crops. The module will be linked to the main FACS2 module through linked dlls that will be separately upgradeable and maintainable, following object-oriented programming style and structure. The modules will comprise diagnostic and predictive phases for each of the establishment, fast growth, and mature phases. Development of the module will follow techniques of 'participatory programming' whereby the potential users critique early prototypes and discuss the software at multiple stages in the development ensuring that the software is appropriate for the intended end user community.