Sampling strategies to estimate field areal evapotranspiration fluxes with a soil water balance approach

Abstract

In this paper, we analyse the spatio-temporal variability of instantaneous and cumulated evapotranspiration (ET) fluxes measured during two long drying periods with a soil water balance approach on 28 point measurements located within a small maize cropped field. Then we test three sampling procedures to estimate field averaged ET and their associated uncertainty. Results of this study show that the ‘within-field’ spatial variability of the local scale instantaneous measured ET is large with coefficients of variation ranging between 15.5 and 46.3%. The corresponding cumulated ET for the first and the second period ranges between 39.4–107.5 and 30.3–55.7 mm, respectively. We show that this variability of ET fluxes can be related in some way to the measured spatially variable crop growth (2.2<Leaf Area Index<4.3). This one has directly an impact on the ratio between the measured actual ET and the reference evapotranspiration (ETo) which generally is smaller than 1 in our field. The first sampling strategy which consists in increasing the number of sampling locations shows that, for a probability level of 95%, the number of samples needed to estimate the field mean cumulated ET flux with a relative precision of 5, 10 and 20% ranges between 40 and 360, 10 and 90, and 2 and 22, respectively. The number of samples required for the estimation of ‘instantaneous’ ET fluxes is even larger. The second sampling strategy which exploits the concept of temporal stability can only be applied for the estimation of cumulated ET. In this case the field averaged ET is estimated with a relative precision of 10% if only one representative location is selected apriori. The third sampling strategy consists of using a covariant such as the LAI or the dry matter yield. For this case, cumulated ET equals 119±6.1 and 118.5±4.5 mm by means of the relationships existing between cumulated ET and LAI and yields, respectively. The latter strategy seems very attractive but necessitates more research to investigate the robustness of the relationships with respect to different climatic and agronomic conditions and to different plant cultivars.