Hi Statisticians/simulation experts, I am doing the following simulations in a fishery problem:
1. Generate a population using a stock-recruitment (S-R) model with a normally distributed random error. 2. Estimate harvest using a catch equation with two input parameters that are varied Uniformly and Normally, respectively, from known values. 3. Estimate 1000 optimum harvest rates for a set of 1000 random errors added to the S-R model (normal variable), parameter 1 in the catch equation (uniform variable), and parameter 2 in the catch equation (normal variable). For example, S-R model+ e1, parameter 1+u1, and parameter 2+e1 produces Optimum harvest rate 1 S-R model+ e2, parameter 1+u2, and parameter 2+e2 produces Optimum harvest rate 2 S-R model+ e3, parameter 1+u3, and parameter 2+e3 produces Optimum harvest rate 3. .................................................................................................................................., etc. Please note that only one set of 1000 normal random errors (e) and one set of 1000 uniform random errors (u) are used to produce 1000 optimum harvest rates. 4. Then, I plot the frequency distribution of optimum harvest rate to get a 95% confidence interval. Is this approach correct? Thanks for any comments or suggestions. Siddeek . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
