Kenya and ENSO: An Observation and La Niña Prediction
Peter E. O. Usher
U N Environment Programme
Nairobi, Kenya
peter.usher@unep.org
Introduction
Awareness of El Niño in Kenya is virtually universal. Eastern Africa has in 1998, suffered unprecedented rainfall resulting in death, disease, food shortages and infrastructure damage totaling millions of dollars which, rightly or wrongly, has been ascribed to El Niño. The Kenyan media have daily reports of events ascribed to El Niño which continue into June, even though the seasons have settled into the normal cold dry season. For example, on 22 June 1998, the Daily Nation reported a plague of rats attributed to conditions brought about by El Niño, of a variety (ratus ratus) said to be as large as cats which are devastating corn fields; and separately the appearance of hippopotamus in a Nairobi city stream swollen river size following months of torrential rain. In delivering his annual budget statement in parliament, Kenya’s Finance Minister blamed El Niño for the nation’s current financial plight; an interesting departure from the perennial culprits, corruption and mismanagement.
The question is - was the unusual and unanticipated rainfall entirely El Niño related and will the future La Niña have an influence on Kenya’s weather?
Kenya’s Climatology
Kenya has four distinct seasons:
Mid-December to mid-March |
hot dry season |
Mid-March to mid-June |
long rainy season |
Mid-June to mid-October |
cool dry season |
Mid-October to mid-December |
short rainy season |
As Kenya straddles the equator, the rainy seasons reflect the biannual passage of the Intertropical Convergence Zone (ITCZ) across the country - southward early in the year and northward in the second half of the year.
The dry seasons coincide with the Indian Ocean monsoon flows: warm during the Northern Hemisphere northeast monsoon and cool during the Southern Hemisphere southeast monsoon reflecting the different origins of the airmasses, their fetch over ocean or desert, whether the air is descending or lifted or accelerating or decelerating.
No wet season is entirely wet, or dry season entirely dry. Significant rainfall can, and has occurred, during any month and rainy seasons may fail almost totally in some locations. Their onset and cessation may be early or delayed and sudden or gradual. The reasons for this large variability from year to year occurs because of the many influences on Kenya’s weather - extratropical as well as local; oceanic and continental; the symmetry or otherwise of meteorological systems north and south of the equator; the vertical profile of the dominant air masses and the modifying influence of mountains and lakes, particularly Lake Victoria.
Kenya and El Niño
Awareness of El Niño first arose from the experience of the 1982/83 event. The event was characterized by dry or drought conditions in the Sahel and in Southern Africa. The plight of those in the Horn of Africa and the farmlands of Zambia and Zimbabwe became familiar to Kenyan through the media. Kenya was also locally dry. The central highlands had average total rainfall. However, the long rains of 1983 finished a month early and the short rains did not arrive until December significantly disrupting planting and harvest and reducing crop yields. December rainfall brought to a close seven months of relatively dry weather, but not enough rain to compensate for a further nine months of drought beginning in January 1984, which included a complete failure of the March to June long rains. The 1985 short rains were wet but 1985 overall was a very dry year.
An examination of all El Niño periods from the present back to 1945, indicate neither noticeable drought nor prolonged rainfall except for the current 1997/98 event which has coincided with a uniquely wet period. Effectively, it has been continuously wet from early October 1997 until mid-June 1998. There was no hot dry season. In fact, record rainfall occurred in many provinces during January and February with Nairobi experiencing the heaviest rainfall for any month since records began with the exception of November 1961 which, at 23.05 inches, was wetter by 1 inch. However, November is climatically a wet month within the short rain season with average rainfall of 6.75 inches; February falls within the hot, dry season and has an average rainfall of only 2.5 inches. Although February 1998 rainfall was uniquely high for the season, wet Januarys and Februarys are not unknown. 1993, for instance, had in January/February 9 inches and 6 inches, respectively, and is notable for a weather-induced railway disaster which cost more than one hundred lives. There have been only seven wet Februarys and four wet Januarys since 1945. Only in 1958 did a wet February coincide with an El Niño year and only in 1961/62 has there been more than three consecutive months of heavy rain (October 1961 - January 1962), again not an El Niño period.
Kenya and La Niña
If no clear relationship can be found between El Niño and climatic variability in Kenya, almost the same can be said of La Niña. Of the seven cold phases since 1945, there have been dry and wet years coincident with La Niña although mostly they have been dry, noticeably so in 1949/50; 1954/56 and in 1984. However, none of these years was as dry as 1976, a non-La Niña year, while the cold episode phase from 1988 to 1990 was a very wet three-year period.
The wettest years on record - 1961, 1963, 1977, and 1998 - were neither El Niño nor La Niña years. The driest years recorded are: 1984, 1969, 1954, 1953 and 1949. Some of these were La Niña years.
Public Perception of ENSO
Before 1998, few people in Kenya were familiar with or interested in El Niño. Farmers were the most sensitive to the issue and it is my observation, that those acquainted with the 1982/83 and 1987 events generally considered the overall effect to be one of dryness, particularly in the Eastern and Central Highlands and the Rift Valley. The coastal and lake regions were on the other hand wetter than normal. There is considerable spacial and temporal variability in the country’s weather at any time and the climatic statistics do not always reveal the vagaries in precipitation distribution. Individual perception is based on personal experience which might not be representative of a broader community or region. Thus, a major Kenyan tea company decided against replanting tea lost to drought prior to the 1997/98 El Niño in the entirely incorrect expectation that the period would mimic the 1982/83 and 1987 situations on the Tea Estate and continue dry.
The Kenya Meteorological Department’s long-range predictions were that early 1997 short rains would continue into the dry season. Although accurate qualitatively, the forecast totally failed to anticipate the quantity of rainfall that occurred, nor did it prepare people for the consequential destruction and economic devastation brought about by the rain which included prolonged closure of the only coast-to-interior highway, vital for the transit of virtually all goods to and from Central Africa as well as the Kenya interior. Rail services were also suspended, due to landslides and neither road nor rail service are likely to regain quickly their pre-rain condition, which was in any case, not particularly good. Currently, bridges are temporary structures, and metalled highways have become cratered, muddy tracks littered with overturned and broken vehicles.
The official forecast did not initially anticipate continuing significant rain throughout the April/May 1998 Long Rains period. In fact, a drought was predicted, whereas May in particular, turned out to be significantly wet and comparable to the previous 1986, 1981, 1980, 1967 and 1958 years with above average May rainfall. Later forecasts were revised to accurately predict the arrival of the cool dry weather appropriate to the season.
The Imminent La Niña
It is likely that the next El Niño will be anticipated with the expectation that it will be comparable with the current event, and devastatingly wet. However, a review of all El Niño events in the last sixty years suggests that it could as equally be dry as wet and with only minimal departure from average conditions. Based on an examination of all cold events over sixty years, it is probable the imminent La Niña will be dry, although this cannot be predicted with absolute certainty. A clue to what might occur might be found in the prevailing situation in Southern Africa. The May rainfall has been below 25 percent of seasonal average in a broad band from Namibia in the West to Mozambique in the East. It suggests that the ITCZ is adopting a more benign character on its northward passage than it possessed on its earlier southward progression from Central Africa. While the character of the ITCZ can alter during its passage north or south, observation suggest that it is more likely to mimic its profile set at its tropical limits. Hindsight suggests it might have been appropriate to acknowledge the highly active nature of the ITCZ in the Horn of Africa in mid-1997, and to have at least considered the possibility of an above average and even exceptional rainy season in Kenya.
As alluded to earlier, there are many factors which contribute to Kenya’s weather patterns. ENSO is one of these and under the majority of situations is unlikely to be the dominant factor. El Niño may or may not have caused the remarkable weather of 1998, although given the abundance of El Niño-induced weather globally, it would take a brave person to deny the connection. Yet heavy rainfall has occurred in non-El Niño years, particularly in 1961/62. The year most similar to 1998 has been 1993 another non-El Niño year. Rather than concentrating on identifying the link between ENSO and Kenya weather anomalies, researchers should also consider the meteorological similarities that might exist during years with comparable weather patterns irrespective of the existence of ENSO.
Conventional wisdom suggests that 1999 might be dry. La Niña events have been coincident with drier periods in Kenya. Public expectation would be for weather of an opposite kind to the 1998 wet ‘El Niño’. If this comes to pass, then it will establish or reinforce a perception of the next El Niño being wet. Policy decisions, if based on this, might be inappropriate and expensive.
It is essential that new investigations of teleconnections are initiated, particularly for those areas where the ENSO signal is more capricious than for those areas where the response is automatic and relatively predictable. Public information on ENSO will also need to be carefully tailored to take account of the many uncertainties surrounding the issue and with our limited ability to make accurate predictions.