Response to MARAM/IWS/2019/PENG/P6.

Abstract

This document2 first corrects some misleading results in PENG/5 through specifying the Operating Models OMs more correctly by including process error explicitly. Results from these revised OMs provide resolution of the “self-test” concerns raised in PENG/P6. However, the negative bias in estimates of the precision of the effect of fishing parameter δ remain unless the magnitude of process error is minimal compared to observation error. Since earlier analyses have indicated that process error dominates observation error in the island closure experiment penguin response data, the possibility remains of large negative bias in the estimates of precision from Sherley et al. models of the effect of fishing parameter based on the use of individual data. Ultimately only simulation tests will reveal definitively whether or not these random effects approaches do improve estimation precision, and it is pleasing to note that the authors of PENG/P6 are now engaged in pursuing such tests.