MEASUREMENTS AND MATHEMATICAL MODELING OF WATER FLUXES IN AN AGRO-ECOSYSTEM OF MAIZE.

During the agricultural seasons 2010 and 2011, monitoring work was carried out in maize field in the experimental farm “A. Menozzi” (Landriano-Pavia) of the University of Milan. Measurements of water balance components on field scale (soil humidity, evapotranspiration, soil water potential, groundwater table depth etc.) were used to build enriched database to be utilized in this study. Water retention characteristics are crucial input parameters in modeling of water flow and solute transport. Several field methods, laboratory methods and theoretical models for such determinations exist, each having their own limitations and advantages (Stephens, 1994). Comparisons between estimated, field and laboratory results is vital to test their validity under different conditions. This study attempts to evaluate and compare those methods. The soil water retention characteristics were determined in two representative sites (PMI-1 and PMI-5) located in Landriano field, in Lombardy region, northern Italy. In the laboratory, values of both volumetric water content (θ) and soil water matric potential (h) were measured using the tensiometric box and pressure plate apparatus. In field soil water content was measured with SENTEK probes, and matric potential with tensiometers. The retention curve characteristics were determined by common and recent PTFs that use soil properties such as particle-size distribution (sand, silt, and clay content), organic matter or organic carbon content, and dry bulk density. Field methods are more representative than laboratory and estimation methods for determining water retention characteristics (Marion et al., 1996). Therefore, field retention curves were compared against retention curves obtained from laboratory measurements and PTFs estimations using root mean square error (RMSE) and bias. The laboratory measurements showed the highest ranking for the validation indices. The second best technique was the PTF Rosetta (Schaap et al. 2001). The lowest prediction accuracy was observed for the Rawls and Brakensiek (1985) PTF which is in contradiction with previous finding (Calzolari et al., 2001), showing that this function is well representing the retention characteristics of the area. We conclude that the Rosetta PTF developed by Schaap et al (2001) appears to be well suited to predict the soil moisture retention curve from easily available soil properties. Modeling water dynamics in the Soil-Plant-Atmosphere (SPA) continuum is an important aspect of crop water management and water transfer through the unsaturated zone is a key hydrological process connecting atmosphere, surface water and groundwater. Many simulation models of water dynamics in the SPA and/or soil crop system have already been developed. Therefore, the use of simplified agro-hydrological models may represent a useful and simple tool to simulate water fluxes VI in the SPA. A physically based approach model SWAP (Van Dam et al., 1997) and a conceptual model IDRAGRA (Gandolfi et al., 2011) were selected to evaluate their performance in simulating water balance components such as soil water content (SWC) in the root zone and actual evapotranspiration (ETact). IDRAGRA is a novel model that has been developed recently in the Engineering Department of the University of Milan, and was compared to a more standardized widely used model (SWAP). Data on various agronomic aspects required for IDRAGRA and SWAP were collected during 2011 growing season of spring-summer maize in two representative sites of the field (PMI-1 and PMI-5). PTF-Rosetta soil hydraulic parameters were fed to the two models intending to compare them with cost-effect, affordable work, and less sophisticated methodology to identify the soil characteristics.. SWAP showed a good performance in estimating ETact and SWC in the root zone in the two sites. Similarly, IDRAGRA showed good fitting with measured data. Then the two tested models were run to assess capillary rise contribution to satisfy maize water requirement and groundwater recharge. When considering the daily variation of the groundwater table, groundwater contribution is able to meet from 47 to 60% for IDRAGRA and from 40 to 65% for SWAP of total maize water requirement. Finally, groundwater table depths were fixed virtually at 0.8m, 1m, 1.5m, 2m and 3m. It was found that the capillary rise from groundwater decreases with the increase of the groundwater table depth. A higher contribution is observed when the water table is higher or equal to 1 m. When the water table depth reached 2m the capillary was still contributing to maize water requirement for both models The deep percolation estimated by SWAP was much higher than the one estimated by IDRAGRA. Accordingly, the net recharge to the water table was considerable with SWAP simulations. While for IDRAGRA model the deep percolation was always lower than the capillary rise which lead to a negligible net recharge Taking all together, a simpler model like IDRAGRA can provide comparable performances to a more complex model, like SWAP, and could therefore be a good tool for estimating water balance components for practical applications, especially for irrigation management.