*2.2. Hydro-Climatic Data*

In general, open access data for the study region is limited in space and time. Long-term observations of the atmospheric variables surface temperature ( *T*) and *P* are readily available from global gridded databases like the CRU TS2.1 data set [26] and the more recently updated CRU TS3.10/CRU TS3.10.01 data sets at 0.5◦ grid-cell resolution [27]. However, for the water variables *R* and *ET* in the landscape, the spatio-temporal coverage of available regional data is much more limited.

Openly reported data in the Global Runoff Data Centre (GRDC) [28] show only three stream discharge stations in Greece with reasonable long-term annual average values for *R*. The reasonableness is judged by the long-term average *R* being smaller than the long-term average *P*, since equal or greater long-term average *R* (as found for other Greek discharge stations) implies unrealistic zero or negative long-term average *ET* = *P* − *R*. The reasons for the physically unreasonable discharge values reported for many stations in Greece are not known to us. For most of these discharge stations and the associated catchments, the annual aggregation of the *R* values (discharge divided by catchment area) implied by the reported discharge data in GRDC equals more or less the annual aggregation of the corresponding catchment-average *P* values in the CRU data; this indicates that the reported discharge data may not be independently measured but derived from catchment-average *P* data without accounting for the partitioning of the *P* water input between *ET* and *R*. At any rate, the three discharge stations with physically reasonable data and their local catchments are located in mainland Greece with the associated rivers being: Mesta-Nestos, Almopaios and Aliakmon (Figure 1). Data from these stations in GRDC [28] have a maximum temporal extent of 24 successive years in the second half of the 20th century (Table 1).

For the Peloponnese, estimates representative of current long-term average *R* conditions are available from reports by the Greek Water Management Authority [29] for the three water managemen<sup>t</sup> districts and associated local catchments NP, WP and EP (Figure 1). For these, average values of *P*, *R* and *ET* are reported for different subcatchments, leading to the average *ET*/*P* ratios listed in Table 1.



the GRDC [28]. d Monthly runoff data available for the period 1966–1989. e Monthly runoff data available for the period 1989–1994. f Monthly runoff data available for the period 1963–1987. g *ET* = *P* − *R*. h Greek Water Management Authority [29]. i Area-average of reported values from different subunits in management districts (as reported from the Greek Water Management Authority [29]). j Considered subunits: GR27 and GR28 (without the island subunit GR45) [29]. k Considered subunits: GR29 and GR32 [29]. l Considered subunits: GR30, GR31 and GR33 [29]. m Effective precipitation partitioning into the reported *R* and *ET* in the district: *Peff* = *R* + *ET*.

### *2.3. Land- and Water-Use Data for Irrigation*

Openly available information on irrigation rates and areas for Greece is not complete over either the whole investigated geographical space or the whole investigated time period. Two different types of such data are available: national irrigation data sets from statistics reported by the Organisation for Economic Co-operation and Development (OECD) [30,31]; and the global map of irrigated areas version 5.0 (GMIAv5.0) [32]. The national OECD data sets [30,31] contain statistics in terms of area that is actually irrigated (*Aai*) and total volume of water withdrawal for irrigation (*Iwwv*) in Greece, especially from 1990 onwards (Figure 2). For earlier times, sporadic data on total irrigated area is provided by [33]. Furthermore, the global map of irrigation areas [32] represents the situation around year 2005 with spatial resolution of 5 arc minute by 5 arc minute (see Figure 3 for Greece).

**Figure 2.** (**a**) Reported irrigated area, *Aai*, in Greece, shown with the best power-law fit to the reported data. (**b**) Reported annual water withdrawal for irrigation, *Iwwv*, in Greece, shown with the best power-law fit to the reported data until year 2001, after which *Iwwv* conditions stabilize. (**c**) Irrigation water withdrawal per irrigated area, *Iww* = *Iwwv*/*Aai*, estimated from the reported and fitted data on *Aai* and *Iwwv* in panels (**<sup>a</sup>**,**b**), respectively. The panels also show average variable values over the two 20-year sub-periods 1930–1949 and 1990–2009.

**Figure 3.** (**a**) Area actually irrigated within each grid cell in km2. (**b**) Irrigated area relative to total cell area. These maps are created from the global map of irrigation areas [32] and represent conditions in the recent period 1990–2009. In these maps, the Ionian and Aegean catchments are indicated by black outlines and the three local catchments in mainland Greece with reasonable runoff data time series are indicated by blue outlines.
