**4. Conclusions**

This paper has developed a general approach for synthesizing and jointly interpreting various types of data, related to different interacting and co-evolving drivers and aspects of freshwater change, subject to various degrees of uncertainty, over hydrological catchments of different scales. The application of this approach to the high-uncertainty case of Greece and its freshwater changes from 1930–1949 to 1990–2009, based on catchment water balance, provides support for an overall decrease in the combined uncertainties of resulting catchment-scale *ET* and *R* changes from the magnitude of the various underlying uncertainties.

For Greece, the study results are robust in showing that climate-driven *P* decrease and concurrent human-driven irrigation development from the first half of the 20th century to recent time have combined in driving considerable total decrease in *R* over the country. In the three largest study catchments (total regional, Mainland and Aegean), the total *R* decrease ranges from −65 to −75 mm/year ±13%. This total *R* decrease is driven by the local irrigation development to more or less the same degree (possibly somewhat more but within the uncertainty range) as by climate change (predominantly decrease in *P*, with relatively small *T* increase over Greece and even decrease in Peloponnese). For the relatively small Ionian and Peloponnese catchments, the *R* decrease is up to −119 mm/year ±15% and −91 mm/year ±17%, respectively, due to the particularly large *P* decrease experienced in these catchments, combined with similar irrigation-driven *R* decrease as in the other catchments.

In each catchment, the irrigation-driven component of the total *R* decrease has largely fed a corresponding irrigation-driven component of *ET* increase. Recent results for water deficit propagation and partitioning over different parts of Europe [16] indicate that this *ET* increase likely represents increased transpiration, thereby contributing to support vegetation and crop status against the decreased *P* water input and the likely more frequent and/or more anomalous drought events that may be expected under such *P* decrease in Greece. Under these conditions of decreased *P* combined with relatively small *T* increase (or even *T* decrease), the atmospheric climate change drives *ET* decrease, that is, opposite change direction to the *ET* increase driven by irrigation. In net total, the irrigation-driven component dominates, such that total *ET* has increased (but less than *R* has decreased) over Greece, except in Peloponnese where the climate—and irrigation-driven components are more or less equal and total *ET* remains essentially unchanged.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2073-4441/10/11/1645/s1, Figure S1: Location and extent of CRU cells considered for each study catchment, Table S1: Base case and alternative (Alt.) quantification scenarios and their variations for different catchments, Table S2: Freshwater changes and main uncertainty estimates for the Mainland catchment, Table S3: Freshwater changes and main uncertainty estimates for the Peloponnese catchment, Table S4: Freshwater changes and main uncertainty estimates for the Ionian catchment, Table S5: Freshwater changes and main uncertainty estimates for the Aegean catchment, Table S6: Evapotranspiration conditions and changes showing insignificant influence of choice of ET/P scenario (defined in main Table 3) for the example of the total regional catchment.

**Author Contributions:** G.D. conceived the study and analysis approach and wrote the paper. C.P. performed most analyses for the Figures and Tables and contributed to the writing.

**Funding:** This research was funded by The Swedish Research Council Formas, gran<sup>t</sup> number 2016-02045.

**Conflicts of Interest:** The authors declare no conflict of interest.
