Search

Evapotranspiration (ET) and net ecosystem exchange (NEE) of a semi-arid grassland (46.69˚N, 19.60˚ E near Bugacpuszta, Hungary) have been measured since 2002 by eddy covariance(EC) technique. The calculation of the turbulent fluxes from the raw (10 Hz) dataset was performed by theEddyPro® open source software. Gap-filling and flux partitioning was accomplished by theREddy ProcWeb online tool (Reichstein et al 2005). Based on the 14 years (2003-2017) long dataset the annual mean evapotranspiration at theBugac site is 488 mm (SD: 77 mm), while the mean annual precipitation for the same periodis 566 mm. The annual sum of ET was less than the annual sum of precipitation except fortwo years (2011, 2012) following a wet extreme year (2010) showing a significant memoryeffect. Mitigating effect of water storage was significant on shorter time scales as well, sinceprecipitation sums during the summer months were exceeded by those of ET indrought/heatwave years (2003, 2007, 2009, 2011, 2012). Strong coupling betweentranspiration and photosynthesis as shown by the strong correlation (R2=0.8618) between theannual sum of ET and GPP was due to high LAI (closed canopy) and more or less constantcover. EC based evapotranspiration data was compared to MODIS Evapotranspiration(MOD16). On annual basis MODIS ET is a close estimation of EC based ET, as the slope ofthe regression between the yearly sums (2003-2013) of MODIS based vs EC based ET is1.0168 (R2=0.6236, RMSE=43.03mm). However, when considering the main growingseason (1 May - 31 July) only, MODIS ET underestimated EC based ET by 10% (R2=0.77,RMSE=22.74), showing that there is a bias on seasonal basis in the MODIS data. Goodnessof fit for the regression of 8-days sums (temporal resolution of MODIS data) in the differentyears range between 0.6913 and 0.8879, while the slope ranges between 0.9271 and 1.2224. [ABSTRACT FROM AUTHOR]