Evapotranspiration for Ireland (ET4I): Ground-Truthing Satellite-Driven Evapotranspiration Products.

Abstract

Evapotranspiration (ET) has been recognized as one of the largest yet most uncertain component of the agricultural water balance and the surface water balance simulated by land surface models. Ireland’s National Meteorological Service (Met Éireann) currently produces 1 km gridded products of rainfall and temperature based on climatological observations. These would be greatly complemented if the potential evapotranspiration (ETo) and actual evapotranspiration (ETa) could be estimated on this grid to provide input into hydrological models and agricultural decision support systems on a daily time scale. Due to the limited availability of observed ET data and the heterogeneous aspect of land use in Ireland, it is difficult to use a statistical interpolation approach to produce gridded maps of ET. Instead, satellite-derived ET products are commonly used, which use remote sensing data to estimate ET at different temporal and spatial resolutions. Open-access satellite ET products that cover the region of Ireland include MOD16, ECOSTRESS PT-JPL, GLEAM, SSEBop, BESS, and WaPOR V3. Utilizing these satellite-derived ET products serves as a clear initial step for the development of a national gridded ET product. However, assessing the accuracy of these products is a prerequisite. While satellite-derived products have been shown to have good agreement with field observations in some regions like Africa, clouded regions such as Ireland are more challenging. As most satellite-based models used to derive ET are based on data in the visible spectrum and involve interpolations between observations, extensive cloud cover in Ireland results in few cloud-free days, leading to inaccuracies and prolonged periods of interpolation. This poster will present the results from a systematic evaluation of these satellite ET products by comparing them to field measurements from flux towers and lysimeters, using a variety of evaluation metrics. Daily ET data from lysimeters and flux towers will be compared to pixel data extracted from satellite ET products at corresponding locations and on the available dates, depending on each product. The flux tower data to be used in ground-truthing is available for a number of sites across Ireland. The temporal range of data availability varies for different sites, with the earliest analysis starting from 2002. Lysimeter data from different locations will also be used for ground-truthing. Analysis using lysimeter data starts from 1990. Some of this data was previously unavailable for analysis, as it existed only in paper records. This project has performed the data rescue necessary to digitize the lysimeter data. This data comparison allows the spatial and temporal accuracy of satellite ET products to be quantified. Additionally, it identifies gaps and limitations in these ET products for Ireland and proposes avenues for refining and advancing ET mapping techniques. This includes the exploration of innovative approaches, such as the integration of machine learning techniques with satellite images and field observations. The study is part of the broader framework of the Evapotranspiration Maps for Ireland (ET4I) project, which aims to highlight the need for enhanced ET modeling and the development of high-resolution gridded daily ET maps for Ireland.