Experimental Seasonal Forecasts (NMME)
The Experimental Seasonal Forecasts presented here are derived from the North American Multi-Model Ensemble (NMME) project (Kirtman et al., 2014) and downloaded from the IRI data library. The NMME project includes 9 models, resulting in a total of 118 ensemble members. Previous studies have found that multi-model ensembles (MME) have higher skill at forecasting climate, and allow for better characterization of prediction uncertainty (Kirtman et al.m 2014, Shukla et al., 2016). The Climate Hazards Center converts these forecasts into seasonal percentile/SPI and then maps them to allow for easy visualization and application.
The NMME forecasts are available between the 5th to 10th of a given month. They are available at 1 degree X 1 degree spatial resolution and monthly temporal resolution. Maps are available for the following regions: Global, Africa, East Africa, Southern Africa, West Africa, Central America, Southeast Asia, and Central Asia. The maps show the probability of seasonal precipitation or temperature for the next 6 seasons (3-month seasons) starting from the month when the forecasts are released (e.g., starting from March-May when the forecasts are released in March).
Seasonal precipitation forecasts probability is presented in terms of percentile-based categories (scale 0 to 100), and SPI (scale <-3 to >+3), and seasonal temperature forecasts probability is presented in terms of percentile. GAMMA distribution is used for seasonal precipitation, and normal distribution is used for temperature.
For each type of normalization (percentile or SPI) there are three categories of maps:
Tercile category maps*: These maps show the probability of seasonal forecasts being Below normal (<33 percentile or <-0.44 SPI), Normal (>=33 to <=67 percentile, -0.44 to 0.44 SPI), or Above Normal (>67 percentile, > 0.44 SPI).
*Climatological probability for these categories is ~33
Moderate category maps*: These maps show the probability of seasonal forecasts being <20 percentile <-0.84 SPI or >80 percentile >0.84 SPI
*Climatological probability for these categories is ~20
Severe category maps*: These maps show the probability of seasonal forecasts being <10 percentile or >90 percentile
*Climatological probability for these categories is ~10
For any given season, the regions shown in white are masked because they are either climatologically dry in that season (for precipitation plots) or because the forecast probability for those regions are not high enough for any of the categories shown (also known as “Equal Chance” regions in the case of Tercile Category maps).
Global
SPI
Precipitation Percentile
Temperature Percentile
Africa
SPI
Precipitation Percentile
Temperature Percentile
East Africa
SPI
Precipitation Percentile
Temperature Percentile
Southern Africa
SPI
Precipitation Percentile
Temperature Percentile
West Africa
SPI
Precipitation Percentile
Temperature Percentile
Central America
SPI
Precipitation Percentile
Temperature Percentile
Southeast Asia
SPI
Precipitation Percentile
Temperature Percentile
Central Asia
SPI
Precipitation Percentile
Temperature Percentile
Reference:
Kirtman, B.P., D. Min, J.M. Infanti, J.L. Kinter, D.A. Paolino, Q. Zhang, H. van den Dool, S. Saha, M.P. Mendez, E. Becker, P. Peng, P. Tripp, J. Huang, D.G. DeWitt, M.K. Tippett, A.G. Barnston, S. Li, A. Rosati, S.D. Schubert, M. Rienecker, M. Suarez, Z.E. Li, J. Marshak, Y. Lim, J. Tribbia, K. Pegion, W.J. Merryfield, B. Denis, and E.F. Wood, 2014: The North American Multimodel Ensemble: Phase-1 Seasonal-to-Interannual Prediction; Phase-2 toward Developing Intraseasonal Prediction. Bull. Amer. Meteor. Soc., 95, 585–601, https://doi.org/10.1175/BAMS-D-12-00050.1
Shukla, S., Roberts, J., Hoell, A. et al. Assessing North American multimodel ensemble (NMME) seasonal forecast skill to assist in the early warning of anomalous hydrometeorological events over East Africa. Clim Dyn 53, 7411–7427 (2019). https://doi.org/10.1007/s00382-016-3296-z
Acknowledgments:
The Climate Hazards Center acknowledges support from the NASA Harvest Consortium, Award No #80NSSC18M0039, and the support of the Defense Advanced Research Projects Agency (DARPA) World Modelers Program under Army Research Office (ARO) prime contract no. W911NF-18-1-0018. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the position or the policy of the Government or the Prime Contract (DARPA and ARO), and no such official endorsement by either should be inferred.