Why Did a Pink Shrimp Landings-Temperature Model Seem to Fail?

Ecological forecasts link effects of physical, chemical, and biological changes on ecosystems (including people) and represent the transition of ecological research into integrated applications.  A landings-temperature model for North Carolina spring pink shrimp Farfantepenaeus duorarum landings that explained 80.3% of the variation in landings during 1962-1991 was used to examine the feasibility of producing ecological forecasts during 2003-2006.  The demonstration was unsuccessful and ended after 4 years because the forecasts overestimated actual landings by 2 to 4 times.  Reduced fishing effort and a change in data collection methodology from voluntary dealer interviews to mandatory dealer self-reports probably affected the model’s reliability.  Even though reported landings were substantially lower than expected after 2000, landings were still directly correlated with winter temperature.  Precision of a linear regression based only on mandatory dealer reports was lower and coefficients were different than the regression based only on voluntary dealer-interview data.  The revised regression might be useful as the basis for exploring further the development of operational ecological forecasts for North Carolina spring pink shrimp landings that rely on mandatory dealer-reported landings.