http://www.agrimet.ir/article_69413_1f753aa38d785ddda126e4772b34e416.pdf

• Main Authors: H. Najafi, A. R. Massah Bavani, P. Irannejad, A. W. Robertson
• Format: Article
• Language: English
• Published: Iranian Society of Irrigation and Water 2018-09-01
• Series: هواشناسی کشاورزی
• Subjects: Seasonal Forecasting of Temperature; North American Multi-Model Ensemble (NMME); Iran’s river basins
• Online Access: http://www.agrimet.ir/article_69417_7a9ddbceeead91b0185eaec5ab3fe826.pdf
• ISSN: 2345-3419; 2588-6002

The aim of this research is to evaluate the temperature outputs of climate forecasting systems over Iran. The analysis is provided based on Atmosphere-Ocean Coupled General Circulation Models from North America Multi Model Ensemble (NMME). The skill of NMME individual models are evaluated in different initializations, of lead times (0-month, 1-month and 2-month) for October-December (OND), December-February (DJF), and February-April (FMA) target seasons. Temperatures at 2m from Climate Research Unit (CRU) dataset are used as reference observation over 1982-2010. Pearson correlation, Mean Error and Root Mean Squared Error are calculated as deterministic verification criteria for seasonal forecast verification. In addition, Relative Operating Characteristic (ROC) score is calculated as a categorical measure for below-normal and above-normal conditions. The results suggest that correlation between NMME forecasts and CRU is higher in FMA (compared to DJF and OND). CFSv2 has a significant skill in the south of Iran in FMA (correlation ≥ 0.9, ROC≥ 0.7). Spatial pattern of NMME biases is similar in three target seasons. GFDL-FLOR-B01 bias is lowest among all evaluated NMME models. At longer lead times; skill of some models is dropped for forecasting temperature in some river basins in Iran. Given large temperature biases found in NMME individual models, applying Model Output Statistics is recommended. Developing Multi-model Ensemble (MME) can also help to improve seasonal forecasts in Iran’s river basins for agriculture and water resources management applications.