Evapotranspiration in a rice paddy field over 13 crop years

Evapotranspiration (ET) in rice paddy fields has been studied over the past half-century, but despite that long history of rice ET study, not much is known about their inter-annual variation. This study investigates ET in rice paddy fields during the growing season for 13 years (2002-2014) at one of the longest flux monitoring sites in Japan. The cumulative ET estimated for each growing season (120 days after transplanting) was 419 +/- 45 SD (standard deviation) mm/season (3.5 mm d(-1)), which increased to 534 +/- 64 mm/season (4.4 mm d(-1)) when energy balance was forced to close. This study determined that ET was correlated with meteorological conditions defined by potential evaporation (Er), and the inter-annual variability of seasonal ET/E-p (coefficient of variation (CV) = 4.6%) was less than half the variability of ET (CV = 10.7%). For most years, the seasonal ET was reasonably estimated as the product of E-p and the overall average ET/E-p (=0.89), including years that experienced both extremely warm and cold summers. However, the errors in estimation were relatively large for 2008 (which experienced high precipitation) and 2012 (which was characterized by an extremely low leaf area index). Although the seasonal ET/E-p was relatively invariable among years, both the seasonal trends of ET and ET/E-p greatly differed between years. The inter-annual variability in ET/E-p was particularly high at the ripening stage (defined as 91-120 days after transplanting). The ET/E-p at this stage was well correlated to precipitation, and the variation was possibly caused by evaporation from intercepted rainfall. The variation in ET/E-p associated with precipitation at the ripening stage accounted for 21% of the inter-annual variation of seasonal ET/E-p.