AQUATER Software as a DSS for Irrigation Management in Semi-Arid Mediterranean Areas

Irrigation management at district or regional scale can be dealt using ecological process-based models and remote sensing data. Simulation crop models simulate at a certain time step the main biophysical variables determining crop photosynthesis and water consumption rates. The research consists in an integrated approach to combine field data, simulation crop model and remote sensing information. Detailed data sets related to topography, soil, climate and land cover were collected and organized into a Geographic Information System, which is routinely updated with remotely sensed images. The code implementation of these two models allows for an improvement of simulation reliability for the crop types considered in the present study in Mediterranean area. Remote sensing images detected by optical and radar satellite sensors at different spatial scales (from 10 to 50 m) have been collected over the analyzed crop cycles. Therefore, remote sensing information about land use and leaf area index (LAI) are assimilated dynamically by the model, to increase the effectiveness of simulation. The integration of crop and water dynamics models with the updated remote sensing information is a Decision Support Systems, AQUATER software, able to integrate remote sensing images, to estimate crop and soil variables related to drought, and subsequently to assimilate these variables into a simulation model at district scale. The significant final outputs are estimated values of evapotranspiration, plant water status and drought indicators. The present work describes the structure of AQUATER software and reports some application results over 2006 and 2007 cropping seasons in Capitanata, South-East Italy. This region has been divided in simulation units cropped by tomato (Lycopersicon lycopersicum L.), sugar beet (Beta vulgaris L. var. saccharifera) and durum wheat (Triticum durum Desf.). Two types of comparison have been carried out: (i) between some tomato observed and simulated data, and (ii) between “LAI Forcing” and “No LAI Forcing” simulated data. LAI Forcing data have been detected by remote sensing over the crop cycle and over the whole region. The model showed a relevant coherence between observed and simulated data (RRMSE = 32, 55, 30% for above ground biomass, LAI and soil water content, respectively). In the case of the application of the LAI Forcing procedure, since simulated LAI are lower than the observed values, as a consequence, simulation results underestimate.