Drainage assessment of irrigation districts: on the precision and accuracy of four parsimonious models

Abstract

Abstract. In semi-arid irrigated environments, the agricultural drainage is at the heart of three agro-environmental issues: it is an indicator of water productivity, it is the main control to prevent soil salinization and waterlogging problems, and it is related to the health of downstream ecosystems. Crop water balance models combined with subsurface models can be used to estimate the drainage quantities and dynamics at various spatial scales. However, the precision (capacity of a model to fit the observed drainage using site-specific calibration) and accuracy (capacity of a model to approximate observed drainage using default input parameters) of such models have not yet been assessed in irrigated areas. To fill the gap, this study evaluates four parsimonious drainage models based on the combination of two surface models (RU and SAMIR) and two subsurface models (Reservoir and SIDRA) with varying complexity levels: RU-Reservoir, RU-SIDRA, SAMIR-Reservoir, and SAMIR-SIDRA. All models were applied over two sub-basins of the Algerri-Balaguer irrigation district, northeastern Spain, that are equipped with surface and subsurface drains driving the drained water to general outlets where the discharge is continuously monitored. Results show that RU-Reservoir is the most precise (average KGE (Q0.5) of 0.87), followed by SAMIR-Reservoir (average KGE (Q0.5) of 0.79). However, SAMIR-Reservoir is the most accurate model for providing rough drainage estimates using the default input parameters provided in the literature.

Retrieving the irrigation actually applied at district scale: assimilating high-resolution Sentinel-1-derived soil moisture data into a FAO-56-based model

“…The SAMIR model implemented at the field scale allows for a robust and explicit representation of crop ET from the crop coefficient values available for all common crops for various climates in the FAO-56 look up tables. It is hence spatializable over large extents from readily available data (Laluet et al, 2023b). Moreover, the assimilation approach developed by Olivera-Guerra et al ( 2023) is used here to invert two parameters of the SAMIR irrigation module: the irrigation amount (Idose) and the soil moisture threshold (SMthreshold) at which irrigation is triggered.…”