JLE

Science et changements planétaires / Sécheresse

MENU

Prévision des besoins en eau d’irrigation du blé par des modèles combinés ARAR Volume 8, issue 2, Juin 1997

Authors
École supérieure d’agriculture de Mograne, 1121 Mograne, Tunisie, Faculté des sciences agronomiques de Gand, Coupure Links 653, B-9000 Gand, Belgique
  • Page(s) : 123-7
  • Published in: 1997

Four historical series of 10-day values for wheat irrigation water needs were established from recorded rainfall and evapotranspiration. These series concern Tunis, Teboulba, Kasserine and Gabes stations. They correspond to observation periods of 20, 5, 10 and 9 years, respectively. They were automatically adjusted by combined autoregressive-autoregressive (ARAR) models. The identified models were used later to predict 10-day water needs of wheat crops. Short-term forecast results differed from observed values by 12 mm on average. Long-term forecasts were much more accurate.