Testing the temporal and spatial validity of site-specific models derived from airborne remote sensing of phytoplankton

Publisher Website
Google Scholar
Abstract
Validations of predictive models are necessary for the accurate application of remotely sensed imagery within ecological research and fisheries management. Multiple regression models derived from airborne imagery on 16 May 1990 accurately depicted phytoplankton biomass and turbidity within aquaculture impoundments. To examine their temporal validity, the exact models, as well as identical model forms, were fit to similar imagery and in situ data collected on 20 June 1990. None of the exact models for 16 May accurately predicted in situ data on 20 June; however, model forms were robust for describing in situ variables. To examine their spatial validity, identical model forms were fit to in situ data partitioned among phytoplankton composition and biomass. The fit of the models and the contribution of imagery variables to the models varied among in situ variables. Although imagery variables explained all of the observed variability for turbidity, regression tree modeling indicated that a significant proportion of the variability in chlorophyll distribution both among and within impoundments was explained through both imagery variables and phytoplankton biomass. Consequently, universal models for the airborne remote sensing of water-quality variables in systems having distinct optical signatures is unlikely. Rather, robust site-specific models will need to be developed.

This publication has 0 references indexed in Scilit: