Area-Averaged Surface Fluxes in a Semiarid Region with Partly Irrigated Land: Lessons Learned from EFEDA

Abstract

Abstract The European Field Experiment in a Desertification-Threatened Area (EFEDA) provides a comprehensive land surface dataset for a semiarid Mediterranean environment with natural vegetation and cultivated dry and irrigated land. This paper discusses the methods and practical aspects of deriving area-averaged fluxes for a range of areas from the whole EFEDA region to several numerical weather prediction model grid cells (on 10–100-km scales). A time series of grid-scale surface fluxes for the entire observational period of 1 month was obtained from weighted surface averages, using a crop phenology–based land use classification together with a homogenized set of surface observations representative of the four major vegetation classes. The flux-aggregated surface observations were compared with two other approaches to obtain grid-scale fluxes (airborne flux observations and radiosondes in conjunction with a simple mixed-layer model). The area-aggregated fluxes (in particular of latent heat) depend strongly on the location of the area boundaries whenever a significant fraction of irrigated land is present. This result confirms clearly the importance of adequately accounting for tiles of irrigated land in surface schemes and corresponding physiographic databases of large-scale models. A simple way to accommodate for minimum information on the canopy water status is proposed in terms of the distinction of at least two seasonal classes of irrigated crops—one of spring and one of summer growing cycles. The main lesson from this aggregation exercise concerns the role of irrigation. First, this study quantifies the uncertainties in the space–time pattern and its effects on aggregated surface fluxes for the first time on the grounds of observational data. Second, it demonstrates practical ways to accomplish the parameterization of irrigation in flux aggregation schemes, by identifying the key data along with their possible sources and by defining a practical implementation procedure.