I would say that, when there is water stress, the plant or tree has more leaves compared to normal irrigation. So leaf mass fraction correspond to the leaf mass compared to total mass (including roots, stem, etc.).
On the leaf level, the leaf mass fraction can be understood as the leaf dry matter content (LDMC), which is the fraction of leaf dry mass to leaf fresh mass [g DM/g FM]. Ideally you should use the saturated fresh mass for the calculation of LDMC to exclude the influence of changing leaf water status during drought stress. An increase in leaf mass fraction during drought stress could be just that. Not an actual increase in leaf dry mass, but a reduction in leaf fresh mass due to lower leaf water content.
In plant ecophysiology, the term "leaf mass fraction" often is used synonymously with "leaf mass ratio", that is the fraction of a plant's total leaf mass to the entire biomass of the plant. Under drought stress (I suppose that you mean drought stress by saying "water stress"), the amount of a plant's total leaf area often is reduced (thereby, loss of water decreases). Therefore, one can expect that the leaf mass ratio also is reduced under drought stress, in particular because, in addition, many plant species tend to increase their root mass under drought stress (however, there are distinct differences among species). More generally spoken, the leaf mass ratio can be seen as the ratio of the "producing" parts of a plant to its "consuming" parts. I hope this helps a bit.