The analysis of long-term irrigation performance series is a valuable tool to improve irrigation management and efficiency. This work focuses in the assessment of irrigation performance indices along years 1995 to 2008, and the cause-effect relationships with irrigation modernization works taking place in the 4000 ha surface-irrigated La Violada Irrigation District (VID). Irrigation management was poor, as shown by the low mean seasonal irrigation consumptive use coefficient (ICUC = 51%) and the high relative water deficit (RWD = 20%) and drainage fraction (DRF = 54%). April had the poorest irrigation performance because corn (with low water demand in this month) was irrigated to promote its emergence, whereas winter grains (with high water demands in this month) were not fully irrigated in water-scarce years. Corn, highly sensitive to water stress, was the crop with best irrigation performance because it was preferentially irrigated to minimize yield losses. The construction of a new elevated canal that 2 decreased seepage and drainage fractions, the entrance in operation of six internal reservoirs that would increase irrigation scheduling flexibility, and the on-going transformation from surface to sprinkler irrigation systems are critical changes in VID that should lead to improved ICUC, lower RWD and lower DRF. The implications of these modernization works on the conservation of water quantity and quality within and outside VID is further discussed.

9Long-term analysis of hydrologic series in irrigated areas allows identifying the main 10 water balance components, minimizing closing errors and assessing changes in the 11 hydrologic regime. The main water inputs [irrigation (I) and precipitation (P)] and 12 outputs [outflow (Q) and potential (ET c ) crop evapotranspiration] in the 4000-ha La 13Violada irrigation district (VID) (Ebro River Basin, Spain) were measured or estimated 14 from 1995 to 2008. A first-step, simplified water balance assuming steady state 15 conditions (with error ε = I + P -Q -ET c ) showed that inputs were much lower than 16 outputs in all years (average ε = -577 mm/yr or -33% closing error). A second-step, 17 improved water balance with the inclusion of other inputs (municipal waste waters, 18 canal releases and lateral surface runoff) and the estimation of crop's actual 19 evapotranspiration (ET a ) through a daily soil water balance reduced the average closing 20 error to -13%. Since errors were always higher during the irrigated periods, when canals 21 are full of water, a third-step, final water balance considered canal seepage (CS) as an 22 additional input. The change in water storage in the system (∆W) was also included in 23 this step. CS and ∆W were estimated through a monthly soil-aquifer water balance, 24 2 showing that CS was a significant component in VID. With the inclusion of CS and ∆W 1 in the water balance equation, the 1998-2008 annual closing errors were within ±10% of 2 total water outputs. This long-term, sequential water balance analysis in VID was an 3 appropriate approach to accurately identify and quantify the most important water 4 balance components while minimizing water balance closing errors. 5

No abstract