Drought Monitoring

The Drought Monitoring Center in Nairobi (DMCN) was established in 1989 to monitor drought conditions in the Greater Horn of Africa (GHA), a region comprising 10 African countries (Burundi, Djibouti, Eritrea, Ethiopia, Kenya, Rwanda, Somalia, Sudan, Tanzania, and Uganda). Some of the techniques the DMNC uses to monitor drought are described below.

Historical Climatic Data

Drought indices are derived by comparing current observations with historical records. The previous records are usually first standardized and then

Jun-Jul-Aug season, Central Rift Valley, Kenya

1. ,. Ill.l 1. 1.


1 1 I"


M 1

Sep-Oct-Nov 1

season, Cent kill

al Kenya 1 1.


1 |l| 1 1

ri|'l '

1960 1965 1 970 1975 1980 1985 1990 1995 2000 2005

1960 1965 1 970 1975 1980 1985 1990 1995 2000 2005

ranked in ascending or descending order. Statistical techniques are used to cluster the current observation within a group of similar past occurrences. The methods that have been used to derive such clusters range from highly sophisticated models like the Palmer drought index (PDI; Palmer, 1965; WMO, 2000) to simple indices such as the tertiles, quartiles, deciles, and percentiles. Most of the drought indices require long-term series of historical data. The choice of an index depends on the availability and extensive-ness of data. For example, the use of percentiles requires about 100 years of data.

Palmer Drought Index The PDI for any individual month (Xi), may be expressed as:

where P is climatically appropriate water balance for the existing condition

Figure 18.2 Annual rainfall distribution in Kenya.

Figure 18.2 Annual rainfall distribution in Kenya.

(mm), Pi is individual monthly precipitation (mm), P is mean annual precipitation (mm), R is soil water recharge (mm), L is soil water loss (mm), PE is annual potential evapotranspiration (mm), and t is number of months. Based on this index, one can characterize the monthly conditions as ranging from extreme wetness (index value > 4) to extreme drought (index value < -4), as shown in table 18.1.

Quartile Drought Index Using this method, historical records are first standardized (based on the specific long-term mean, standard deviation, and sometimes higher order statistics), ranked, and then divided into four groups based on quartiles, as shown in table 18.2. All new observations are then classified into one of these groups based on the magnitude of the specific observations. Assessment of drought severity in the GHA is based on the quartile index. Cumulative monthly rainfall is used to assess the persistence of drought. Figure 18.5 compares the worst drought conditions that occurred during 1984 and 2000 on the basis of cumulative rainfall.

The 2000 drought was associated with a La NiƱa event due to the observed slow cooling of sea-surface temperatures over much of the tropical Indian Ocean. Chapter 3 describes such atmospheric conditions in detail.

Figure 18.3 Arid and semiarid Lands of Kenya (from Ominde, 1971).

Figure 18.3 Arid and semiarid Lands of Kenya (from Ominde, 1971).

Drought indices indicated that the 2000 drought was far more severe than that of 1984 in many parts of Kenya. Many people lost their lives in 1984. Millions were displaced in search of water, food, and grazing land. Although the 2000 drought was more severe, the impacts were much less severe due to the early warning provided by DMCN and Kenya Meteorological Department (KMD).

Remote Sensing

Remotely sensed data can now be used as proxies for records. In the case of agriculture, the NDVI (chapters 5 and 6) is one such proxy for monitoring vegetation conditions, while cold cloud duration (CCD; chapter 20) and cloud temperatures are common proxies for rainfall. A comparison of the near-real time composite NDVI values with the average and any past records can help delineate areas with relatively drier and/or greener vegetation conditions. In Kenya, NDVI has been used to monitor vegetation cover, which together with the quartile indices is used to monitor drought. For example, NDVI images showed a general deterioration in vegetation conditions over northern and eastern Kenya, where some locations experienced the driest conditions since 1961 in May 2001 (DMCN, 2001).

Landsat series has also provided continuous data on the conditions of the earth's terrestrial surface for more than 25 years. These data are crucial in addressing the consequences of drought through ecosystem mapping,

Figure 18.4 Impacts of the 1999-2001 drought on (A) maize and (B) livestock in Kenya.

Figure 18.4 Impacts of the 1999-2001 drought on (A) maize and (B) livestock in Kenya.

Table 18.1 The classification of drought or weather conditons based on Palmer drought index

Index value

Drought classification

> + 4

Extreme wetness

+4 to +3

Severe wetness

+3 to +2

Moderate wetness

+2 to -2

Near normal

-2 to -3

Moderate drought

-3 to -4

Severe drought

< -4

Extreme drought

Source: Palmer (1965).

Source: Palmer (1965).

Table 18.2 The quartile drought severity index (QDSI) used by the Drought Monitoring Center, Nairobi, Kenya





Rainfall < min

Driest on record


Min < rainfall < Q1



Q1 < rainfall < Q3

Near normal


Q3 < rainfall < max



Rainfall > max

Wettest on record

Source: Ogallo (2000).

Source: Ogallo (2000).

deforestation, land-cover change, and forest and grassland fires. Many other remotely sensed data that can be used in drought monitoring can be obtained from USGCRP (1999) and WMO (2000).

Livestock Conditions

The condition of livestock (camels, cattle, goats, and sheep) can also be used to monitor drought conditions. The vegetation density of the Kenya's arid and semiarid lands is low, and the variation in forage quantity and quality is enormous. These variations and periodic lack of water for livestock due to low rainfall force pastoral communities (e.g., Maasai, Sam-buru, Turkana, and many others) to wander continuously with their herds of livestock. Due to drought, density as well as quality of pasture deteriorates and so do the physical conditions of the livestock that feed on these pastures. Many livestock get weaker and weaker and eventually die of starvation (figure 18.4).

Low livestock prices are usually signs of drought stress when they are being sold for money to purchase expensive food commodities or when destocking is taking place to avoid eminent death due to lack of water and forage. Livestock numbers and prices when properly monitored would give an indication of food security in a particular locality. The meat-cereal price

Lodwar station, North West Kenya

Figure 18.5 The cumulative

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Lodwar station, North West Kenya

Figure 18.5 The cumulative

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec monthly rainfall during the worst drought conditions during 1984 and 2000 in Kenya.

ratio can also be used to monitor drought. During drought, livestock are sold cheaply, and the meat price declines while the price of cereals (such as maize) increase.

The destruction of crops and livestock reduce the economic status of most rural communities. Some of the communities in marginal areas, for example, resort to survival mechanisms such as reducing the number of meals per day during periods of food shortage. Others participate in deforestation for the purpose of charcoal burning to meet income deficits (Karanja et al., 2001). Dietary/food composition also changes during drought conditions. Cereal intake by humans, which is usually higher during nondrought years, declines, and consumption of pulses (relatively drought resistant) increases during drought years.

Drought Early Warning System

The DMCN is an intergovernmental center for climate monitoring for the GHA. The DMCN and KMD play an important role in providing weather and climate advisories, including predictions and early warnings about severe climate events such as droughts. Using statistical methods, weather forecasts for 10-day, monthly, and seasonal durations are provided by the DMCN and KMD as part of their normal operation activities. (DMCN, 2001).

0 0

Post a comment