How the platform works

The MOSES platform will be implemented according to the most recent enterprise IT and web-computing frameworks, in compliance with geospatial standards such as W3C, OGC, ISO, INSPIRE, allowing to merge multi-source information and helping to scale-up. The main functionalities of the MOSES platform (see Figure below) are:

Context diagram of the MOSES platform

Context diagram of the MOSES platform

  1. Seasonal probabilistic forecasting / downscaling. This module provides probabilistic forecasts produced by meteorological centres. Seasonal forecasting in Europe is currently based on multi-model ensemble technique available at ECMWF within the “EUROSIP” system: http://www.ecmwf.int/products/catalogue/VIII.html. Different techniques to calibrate and combine model outputs produced by EUROSIP centres and other Global Producing Centres will be applied to optimize the quality of the final probabilistic forecasts. Where deemed useful, the direct model output will be regionalized and calibrated on local climatic data by means of statistical downscaling techniques [R14][R15][R21].
  2. Early in-season crop mapping. This module provides agricultural land cover classification maps from multi-temporal satellite optical images and, when available, field data provided by users. The MOSES project will take advantage from the Copernicus Programme by exploiting, when available, optical and radar images from the Sentinel EO space missions. A feasibility study for the service evolution will analyse the capabilities and potentialities of data fusion between SAR and optical data for enhancing and improving early-season crop classification and in-season crop status monitoring.
  3. Long and medium term irrigation forecasting. This module performs, both at long and medium term, irrigation water demand forecasting for areas under MOSES service. The module includes a daily weather data generator and runs the water balance model (CRITERIA 3D [R9], or a similar one to be agreed in the design phase), estimating and scheduling irrigation water consumption. The long term forecasting service is provided once early in the season, starting from seasonal probabilistic climate forecasting and a local climatic dataset; it will support users in the definition of optimal planning for water procurement and allocation before the beginning of the irrigation period. The medium/short term forecasting service is instead provided during the irrigation season starting from medium/short term numerical weather forecasting and in-season monitoring of the evapotranspiration and water storage status. This tool enables the user to perform a dynamic and flexible management of water resources during the season.
  4. In-season water demand monitoring. This module monitors and refines the seasonal crop irrigation demand forecasting, producing a performance analysis useful for assessing water demand and delivery needs during the irrigation season. The innovative module provides in-season monitoring of evapotranspiration and water storage status; this information is compared with the crop irrigation demand forecast, issued at the beginning of the irrigation season, and with the crop tailored daily output of the irrigation model in Italy in order to program day-by-day water resource management activities; it is mainly based on satellite data and related techniques.

During the irrigation season two time series of EO data will be assimilated into the water balance model: 

  • Crop water requirements estimated by generic algorithms without crop specific parameters using multi-spectral intermediate resolution satellite images (Landsat8, SPOT4 and later Sentinel2);
  • Soil moisture content retrieved from active and passive microwave instruments, such as SMOS, SMAP, ASCAT[R10];

The figure below  is a depiction of the timeline of a typical operational scenario of the MOSES system.

The activity timeline of the MOSES system

The activity timeline of the MOSES system

The time arrow covers about one year, and we suppose in the figure that the irrigation season to last about three months, but adaptation to different conditions (up to about eight month) is of course feasible. The process of EO data acquisition for crop mapping starts several months before the beginning of the irrigation season start, and the single seasonal irrigation forecast is released as soon as crop maps and seasonal weather forecasts are available. Afterwards, a monitoring phase starts, requiring the acquisition of updated EO and meteorological data for the production of in-season water demand maps and the refinement (medium and short term) of water irrigation demand forecasts. After the irrigation season, validation of algorithms/models with respect to the ground truth and actual seasonal weather is performed, together with the analysis of obtained and potential water and cost savings.

Go to the Portal