Context :

Aims : We aim to perform a theoretical evaluation of the impact of the mass loss indetermination on asteroseismic grid based estimates of masses , radii , and ages of stars in the red giant branch phase ( RGB ) .

Methods : We adopted the SCEPtER pipeline on a grid spanning the mass range [ 0.8 ; 1.8 ] M _ { \sun } .
As observational constraints , we adopted the star effective temperatures , the metallicity [ Fe/H ] , the average large frequency spacing \Delta \nu, and the frequency of maximum oscillation power \nu _ { max } .
The mass loss was modelled following a Reimers parametrization with the two different efficiencies \eta = 0.4 and \eta = 0.8 .

Results : In the RGB phase , the average random relative error ( owing only to observational uncertainty ) on mass and age estimates is about 8 % and 30 % respectively .
The bias in mass and age estimates caused by the adoption of a wrong mass loss parameter in the recovery is minor for the vast majority of the RGB evolution .
The biases get larger only after the RGB bump .
In the last 2.5 % of the RGB lifetime the error on the mass determination reaches 6.5 % becoming larger than the random error component in this evolutionary phase .
The error on the age estimate amounts to 9 % , that is , equal to the random error uncertainty .
These results are independent of the stellar metallicity [ Fe/H ] in the explored range .

Conclusions : Asteroseismic-based estimates of stellar mass , radius , and age in the RGB phase can be considered mass loss independent within the range ( \eta \in [ 0.0 , 0.8 ] ) as long as the target is in an evolutionary phase preceding the RGB bump .