Estimating the uncertainty of GPP with including the uncertainty in temperature sensitivity E0 requires a bootstrap of the data in each period, which consumes many resources. Currently, this is done for each uStar scenario in sEddyProc_sGLFluxPartitionUStarScens. However, usually only the uncertainty is inspected for only one scenario.
The uncertainty estimate is used to determine, whether parameter are within bounds, but can be computed in a cheaper manner from the Hessian of the fit. Although this is a conservative estimate, that neglects uncertainty in E0, this is sufficient for the check of parameter bounds.
Omit the expensive uncertainty estimate for other uStar scenarios except one user-specified scenario.
Estimating the uncertainty of GPP with including the uncertainty in temperature sensitivity E0 requires a bootstrap of the data in each period, which consumes many resources. Currently, this is done for each uStar scenario in
sEddyProc_sGLFluxPartitionUStarScens. However, usually only the uncertainty is inspected for only one scenario. The uncertainty estimate is used to determine, whether parameter are within bounds, but can be computed in a cheaper manner from the Hessian of the fit. Although this is a conservative estimate, that neglects uncertainty in E0, this is sufficient for the check of parameter bounds.Omit the expensive uncertainty estimate for other uStar scenarios except one user-specified scenario.