The challenge of disseminating the seasonal forecast during the unusual 2012/2013 rainy season

The seasonal forecast of the forthcoming rainy season is of paramount importance, particularly in the SADC counties where rain fed agriculture is the backbone of the economies of many member states in the region. It has become an important tool in many key sectors such as disaster risk reduction, water management and health. Recent developments in science have improved our understanding of the El Niño Southern oscillation (ENSO) which has raised hope for good quality forecasts since ENSO is thought to be one of the primary factors modulating the rainfall over the region. Although the relationship between ENSO and rainfall over Southern Africa is not linear, during El Nino years below-normal-rainfall conditions are more likely to occur whereas during the La Niña years, the sibling of El Niño, wetter than normal conditions are expected.

The 2012 summer was shaping up as déjà vu, after two years of La Niña events an El Niño was on its way. In fact, the majority of climate forecast models were in agreement that weak El Niño would develop in September. As a Meteorologist, I was appointed to come up with the seasonal forecast for Mozambique. Because of the already known relationship between rainfall and El Niño over Southern Africa I was happy and confident to produce the seasonal forecast thinking that it would be business as usual. The elaborated forecast called for elevated chance for below rainfall conditions to occur over Southern Mozambique during October-November-December (OND) and January-February-March (JFM). Because of the ever increasing need for tailored seasonal forecast from the stakeholders I was even bold:

I decided to forecast the onset of the rainy season. The forecast was calling for a slightly increased chance of a late onset of the rains to occur in southern Mozambique.

My soaring confidence was shaken in mid-November when the forecasted El Niño conditions did not materialize. The expected El Niño conditions retreated, giving rise to neutral conditions sometimes referred informally as La Nada (nothing in Spanish). It is an unusual situation in the last 50-60 years.

This turnaround in weather patterns wreaked havoc in Southern Mozambique as floods were recorded which claimed at least 100 lives. Furthermore, in contrast to my forecast, the rains in Mozambique started very early in this season.

Numerous studies have identified six major constraints to seasonal forecast effectiveness:

1. Legitimacy: when users question the political agenda of the communicator
2. Scale: the forecast is coarse and lacks the local information
3. Cognition: users do not understand the language and indigenous knowledge is not included
4. Procedures: there’s a lack of interaction between the forecasters and end-users
5. Choices: is the forecast accurate enough to alter any real decision?
6. Credibility: the failure of past forecasts to be accurate

As one can clearly see from the above constraints, I am struggling with the last one (CREDIBILITY). If users perceive the forecast communicator as having been wrong in the past it is unlikely that the communicator will be trusted.

It seems that my reputation has been eroded. Am I eligible to do the forecast for the next rainy season? It is a big dilemma.

References:

Hansen JW, Mason SJ, Sun L and Tall A 2011: Review of seasonal climate forecasting for agriculture in Sub-Saharan Africa. Expl. Agric, 47(2),205-24

Patt A. and Gwata C. 2002: Effective seasonal climate forecast applications: examining constraints for subsistence farmers in Zimbabwe. Global Environmental Change, 12, 185-195.

Where did that El Nino go? Wiped out by unprecedented cool shift?

La Niña is over, La Nada may be next