Variance - Based Sensitivity Analysis : 1 COAMPS

4 Numerical weather prediction models have a number of parameters whose values 5 are either estimated from empirical data or theoretical calculations. These values are 6 usually then optimized according to some criterion (e.g., minimizing a cost function) in 7 order to obtain superior prediction. To that end, it is useful to know which parameters 8 have an effect on a given forecast quantity, and which do not. Here, we demonstrate 9 a variance-based sensitivity analysis involving 11 parameters in COAMPS R ©.1 Several 10 forecast quantities are examined: 24hr accumulated 1) convective, 2) stable, 3) total 11 precipitation, and 4) snow. The analysis is based on 36 days of 24hr forecasts between 12 Jan. 1 and July 4, 2009. Regarding convective precipitation, not surprisingly, the 13 most influential parameter is found to be the fraction of available precipitation in the 14 Kain-Fritch cumulus parameterization fed back to the grid scale. Stable and total 15 precipitation are most affected by a linear factor that multiplies the surface fluxes; and 16 the parameter that most affects accumulated snow is the microphysics slope intercept 17 parameter for snow. Furthermore, all of the interactions between the parameters are 18 found to be either exceedingly small, or have too much variability (across days and/or 19 parameter values) to be of primary concern. 20 COAMPS is a registered trademark of the Naval Research Laboratory. 2