COLT Project: Real Time Agricultural Classification for Water Needs Estimation in Emilia-Romagna, Italy

Level: National
Region: Europe
Tags: Agriculture | Institutional measure
Target audience: Agricultural authorities | Environmental authorities | Farmers | Scientists | Students/university | Water authorities

The aim of the COLT project, which started in 2007/2008, was to classify a set of satellite images acquired during the first part of the agricultural season and to estimate the water needs of crops using CRITERIA, a soil/water-balance modelling system. The method also relies on a modified version of the Wofost crop growth model and a modified version of the LeachM model. The methodology is based on three to four new satellite images obtained each year between November and June from the DMCii satellite constellation, which take in the whole study area in one pass. The approach also requires use of a sample of field surveys covering around 600 plots that are equally distributed in the focus area. Crop maps are ready between mid-June and mid-July (depending on satellite acquisition times), right at the start of irrigation season (typically peaking in July). They are provided to both the Reclamation Consortia and the Regional Agriculture Department for water management and statistical use.

Irrigation water savings can be achieved through the calibration of the water needs of various crops and fruit orchards. Farmers can obtain information on the water needs of their fields, based on which they can avoid sub-optimal over-irrigation and under-irrigation. Since over-irrigation is more likely to happen (“to be on the safe side”, at least when enough water is available), this methodology contributes to water savings while enhancing yields and lowering irrigation costs.

The results of the project can also be applied to help establish priorities for the allocation of limited water resources among irrigation needs.


  • The COLT project aims to classify a set of remote sensing images acquired during the first part of the agricultural season. The main idea is to classify the agricultural land before the beginning of the irrigation season.
  • The study area, about 1 million hectares, is in the plain of Emilia-Romagna, within the Po and Reno river basins.
  • The methodology is based on three to four new satellite images obtained between November and June from UK-DMC, field surveys, image classification, vectorisation, and water-balance modelling using CRITERIA software. CRITERIA is a soil water balance modelling system that includes a modified version of the Wofost crop growth model and a modified version of the SoilN model as supplementary modules. The results are provided to the Regional Agriculture Department and the Reclamation Consortia to assist with water management-related decision making (e.g. identifying allocation priorities when there is not enough water) and for statistical purposes.
  • The developed methodology was first tested in 2007/2008. Four sets of images were acquired (in November, February, April and May), and by using these high-resolution data sets it was possible to verify the operational situation on the ground. At present, the synthetisation of field data (e.g. crop choice by plot) and satellite images seems adequate for the purpose. The July model of calibration is necessary to properly handle summer crops. The precision of satellite images to determine land use is quite high: 90 percent for the winter crop class, 85 percent for maize in the summer crop class, 70 percent for sugar beets in the summer crop class, 83 percent for alfalfa in the meadow class, and >90 percent for orchards and vineyards in fruit tree class.
  • The test also helped to calibrate the methodology and to achieve the best possible quality and detail at the lowest possible cost.
  • The output data is used to set water distribution priorities according to crop types — i.e. to avoid under-irrigation and over-irrigation. This implies close to optimal use of scarce water resources for seasonal irrigation. The method also helps to estimate water use by farmers, which is the basis of payments for irrigation water; it also helps to assess drought damage.

Results obtained

  • The project established a successful pre-operational service for the “real time” monitoring of crops to define water needs for agriculture via application of an extended water balance model.
  • Extracted data from satellite images are exploited through an operational chain managed by the Reclamation Consortia to aid decision making on a range of issues, such as water distribution priorities according to crop types.

Success factors

  • Advanced technology using satellite images
  • Trust of institutions, farmers and other stakeholder in the technology (which can be enhanced through education)

Indicators used

  • Saved water amount for irrigation
  • Number of hectares covered with the scheme

In principle, the methodology can be applied in any location.

Preconditions for application include acquisition of field data, cooperation among various entities such as agricultural institutes, farmers’ associations, service providers (satellite companies) and water agencies. A high level of organisation between multiple players is essential.

Total costs

  • EUR 40,000 per year