Evaluation of a model for predicting sucrose yields following hurricane damage to sugarcane in Hawaii

A census taken within 3 months after Hurricane Iwa struck the island of Kauai in November 1982 provided data for a model of growth and production from which to calculate an estimated percent yield loss due to the hurricane. The estimated percentage loss was then subtracted from the potential yield projected from a linear regression of the five preceding years to give a modeled yield estimate. The model-estimated yields and the projection-estimated yields of two hurricane-struck plantations for the two years following the hurricane were compared to the actual yields reported by their sugar mills. For the individual plantations over the two years the modeled yields ranged from 0.3% under to 28.4% over the actual yields.The model gave yield estimates closer to the actual recorded yields than did the projected yield estimates, and it was more accurate the first year than the second year following the hurricane. Additionally, the model was more accurate for the rain-fed plantation than for the irrigated-plantation. Thus, the model tested here for evaluating sugarcane yields in response to hurricane damage varied in accuracy by location and year. Nevertheless, the model provides a method to derive quantitative estimates of natural hazards on crop production and is a valuable tool.