Evapotranspiration (ET) is the most significant water balance component and is also a very complex component to evaluate in spatio-temporal scales. Remotely-sensed data greatly increases the accuracy of basin wide ET estimation but only in periods with available satellite images. This paper describes an attempt to estimate daily ET regardless of the availability of the satellite images. The method is based on application of the interpolated evaporative fraction (Λ) from "historical" satellite images to periods with no satellite data available. Basin wide daily ET is obtained by combining interpolated Λ and standard PET methods on meteorological stations. The reliability of such approach was evaluated by comparing the obtained daily ET to the SEBAL ET estimates through the analysis of residuals (∆), standard deviations of residuals (σ) and the Nash-Sutcliffe coefficient (NSE) over the basin. The SEBAL ET estimates were validated with the data from two lysimeters. The discrepancy of obtained ET versus the SEBAL ET estimates (∆ = 0.13 mm day −1 , σ = 0.64 mm day −1 , NSE = 0.07) indicated that the proposed concept has relatively high accuracy, which is notably higher than the Penman-Monteith interpolated ET estimates (∆ = 1.94 mm day −1 , σ = 1.03 mm day −1 , NSE = −4.71). It was shown that a total of five images can provide a reliable estimate of interpolated Λ and thus represent specific characteristics of a basin. As the presented concept requires minimum remote sensing data and ground based inputs, it could be applied to estimate basin wide daily ET in data scarce regions and in periods with no satellite images available.

A comparison of various methods that enable temporally continuous computation of basin-wide air temperature is presented. An approach that combines remote sensing data with measurements at meteorological stations for obtaining basin-wide air temperature is proposed and compared to the standard interpolation methods of point measurements. For a basin of over 1000 km2, the proposed approach provides significantly more reliable air temperature rasters (average Δ = 9%) than the standard interpolation methods (average Δ = 25%), all by using satellite images and measurements from only two meteorological stations in comparison to standard methods using measurements from 10 meteorological stations.