Developing economic thresholds for rice stink bug management in Texas using dynamic programming

Abstract

Updating of the treatment recommendations used by Texas rice producers' for the management of the rice stink bug (RSB) is the objective of this study. Present treatment recommendations used in Texas and throughout the southern rice growing region have not been updated since 1981. These thresholds are static in nature and lack the flexibility to account for changing market conditions, rice yields, and RSB control costs. Flexible economic thresholds for the control of the RSB under Texas conditions were developed using a stochastic dynamic programming model. Data on RSB damage and population dynamics was provided by Dr. M.O. Way for two locations in the Texas Rice Belt (Beaumont and Eagle Lake) for the period 1984 to 1987. Development of the dynamic programming model required five steps. First, important damage relationships were identified and estimated. Second, the stages and state variables describing the system were selected and the Markov requirement for the state variables was deemed to be satisfied. Third, the behavioral objective of profit maximization was selected and operationalized. Fourth, the recursive equations for the dynamic programming model were constructed and the probability distributions for the stochastic state transitions were estimated. Finally, the dynamic programming model was validated to ensure that it operates in the expected manner. Specific thresholds for the 1988 rice growing season were developed using the dynamic programming model. The primary conclusions are: 1) linear damage functions fit the data best, 2) damage functions for peck and head yield with adult RSB at different stages of rice plant development as the independent variables provide the best combination of explanatory power and sampling ease, 3) the level of RSB infestation has no influence on yield, 4) strong Markov relationships exist between the number of adult RSB in the present time period with those in the previous time period, 5) the thresholds decrease as price, first crop rice yield, planting date, and discounts for damage increase, and 6) the thresholds increase as treatment costs and discounts increase. The thresholds are lowest at the milk stage of rice plant development; the thresholds at soft and hard dough are generally higher.