The Available WAter REmaining (AWARE) model highlights the importance of considering consumption rather than withdrawal and takes into account spatial variability. It results from a massive and collective effort on behalf of the Water Use in LCA (WULCA) working group. The AWARE model provides a consensual, operational and recommended indicator for addressing and comparing water impacts, and fully succeeds in this purpose. The present work discusses the shape of the model, as well as associated limitations on its range of validity, which do not distinguish between regions that are more degraded than fair. A subsequent improvement is then proposed. This improvement is based on the ratio of the demand (requested by the ecosystem) to availability (minus the effective human appropriation), more simply named the demand-to-remaining (DTR). It provides useful and straightforward information representing the current state. follows the common practice in LCIA by (1) the definition of a relationship modelling the impact according to human intervention and (2) the use of marginal approach for determining the characterisation factor (CF). The significance of the approach is addressed by the sensitivity of the CFs according to the components of the model. As expected, the CFs of the AWARE and DTR models increase linearly with the area, in the same manner. In both situations, an increase in C_H produces the same result as a decrease in A. The increase grows faster when the AWARE model upper boundary is being reached, and when the complete human appropriation of water (C_H=A) is being attained for the DTR model. This implies that the relationships present similar features but at different intervals, without any discontinuities for DTR. AWARE consensus model brings a major benefit to the community by proposing a shared standard. However, AWARE relationship is only defined when human consumption has spared sufficient water for an ecosystem in fair condition and loses its validity for more severe situations. By defining impact as the fraction of ecosystem demand on what is left by human activity, the DTR model proposed in the present work makes it possible to overcome this limitation. This improvement is mathematically sound, all the while satisfying the same expectations as the AWARE model.