I made some changes to check out the possibility of using the precipitation factor as free variable (and of fixing the melt_f)
created a minimize_bias_geodetic_via_pf_fixed_melt_f, which calibrates the precipitation factor to match the geodetic data when the melt_f is fixed (same as minimize_bias_geodetic, but with pf as free variable!)
created a optimize_std_quot_brentq_geod_via_melt_f, which seeks for the optimal melt_f for WGMS reference glaciers when we use the pf as free tuning parameter
also added a test, where I found that the "allowed melt_f" or optimize_std_quot_brentq_geod_via_melt_f should not be too small, otherwise brent_q can't find an optimal melt_f (already for the HEF, I needed to set the allowed range to melt_f=(150,600) )
also changed the default year range of minimize_bias_geodetic to end of 2019 (because we have to use this anyway if we want to calibrate on geodetic, before it was only set to end of 2018, because no climate data was available till end of 2019...). This should not change anything, because in the scripts we have so far prescribed the year range
I made some changes to check out the possibility of using the precipitation factor as free variable (and of fixing the melt_f)
created a
minimize_bias_geodetic_via_pf_fixed_melt_f, which calibrates the precipitation factor to match the geodetic data when the melt_f is fixed (same asminimize_bias_geodetic, but with pf as free variable!)created a
optimize_std_quot_brentq_geod_via_melt_f, which seeks for the optimal melt_f for WGMS reference glaciers when we use the pf as free tuning parameteralso added a test, where I found that the "allowed melt_f" or
optimize_std_quot_brentq_geod_via_melt_fshould not be too small, otherwise brent_q can't find an optimal melt_f (already for the HEF, I needed to set the allowed range to melt_f=(150,600) )also changed the default year range of
minimize_bias_geodeticto end of 2019 (because we have to use this anyway if we want to calibrate on geodetic, before it was only set to end of 2018, because no climate data was available till end of 2019...). This should not change anything, because in the scripts we have so far prescribed the year range