Thesis colloquium Klaas Thomas Jellema

The set-up of FLUXNET, a world-wide network of Eddy-covariance measurements, gave opportunities to make use of large scale high quality Evapotranspiration data. A space-time analysis has been performed for continental (CarboEuropeIP) and regional (KNMI) (sub-)networks. In this analysis the dominant periodicities of the variogram parameters were determined, though facing several quality problems.

Organisator Hydrology and Quantitative Water Management

vr 31 mei 2013 11:00 tot 12:00

Locatie Lumen, building number 100
Droevendaalsesteeg 3a
6708 PB Wageningen
+31 317 481 700
Zaal/kamer Lumen 1

Investigation of the space-time variability of evapotranspiration data at regional and continental scales

A study between the KNMI and FLUXNET-CarboEuropeIp networks


The increasing tension on water resources due to population growth and future climate change leads to a world-wide call for accurate water management. This is only possible if observations of all components of the water balance can be measured with a sufficient accuracy. Evapotranspiration (ET) is normally considered as a rest term, and therefore often not determined accurately. Relatively new methods as the Eddy covariance technique provide accurate observations of Actual ET rather than Potential ET derived from meteorological observations, a common used method. The set-up of the FLUXNET network, a conglomerate of world-wide measurement sites with Eddy covariance obtained data, gives opportunities to make use of large scale high quality data of ET. A first step to investigate its suitability with respect to current methods is to look at the data availability, quality, and space-time variability. The analyses have been carried out with data from regional (KNMI) and continental (FLUXNET-CarboEuropeIP) networks. The spatial structure of ET was characterized with the sill, the time dependent maximum semi-variance. Based on variogram analyses, the sill shows a clear diurnal and seasonal variation for the different used ET data. With a multi-linear regression analysis this pattern is characterized by fitting a double-cosine function. Nevertheless, the associated length scale, the range, did not show any clear periodicity. Six schemes of x-hour moving average windows were applied to the ET data to ascertain the dependence of spatial patterns for different temporal resolutions. Also in this case only the sill shows a clear diurnal and seasonal variation, though this pattern diminishes strongly with increasing resolution. Moreover, the coefficients of the fitted parameters of the double-cosine function show a strong power-law relationship versus the resolution. In a close inspection of the created variograms instability and quality problems emerged. This asks for more research concerning the accurate determination of variograms in (sub-)daily periods and for different temporal scales. Above all, this research could be seen as one of the few attempts to analyze the space-time variability of ET data at an European scale.