Current Climate in East Africa

Introduction

The climate of East Africa is highly variable in the best of times — droughts especially have been a “normal” phenomenon of the climatology of the region. Recently, however, temperature trends have started to change, and the region’s temperature is already significantly warmer. Rainfall is also changing — some areas are getting drier and others wetter, and variability appears to be increasing. This section will provide some data to illustrate these changes. The subsequent sections discuss impacts.

Climate Data Analysis Methodology

The team has acquired daily and monthly rainfall data for select locations from the governmental Meteorological Departments in Kenya, Uganda and Tanzania. The data was statistically analyzed to identify trends, especially in rainfall.  The quality of the datasets range from being excellent (e.g., Katumani, Kenya) to having too many missing values to permit temporal trend analysis. The analysis was done via a median of pairwise, slopes regression. This is a regression procedure that calculates the slopes between every possible pair of points in the data set. The intercept is calculated from the median of the residuals of the data points vs. the slope (Hoaglin et. al 1983). A nine year moving average is a standard for graphs of rainfall since it smooths  the year-to-year fluctuations but reveals trends, and few climate phases (such as the El Niño Southern Oscillation or ENSO) coincide with that length.

Current Climate Trends

Climatology

Rainfall in the East Africa region is dependent on the movements of the Inter-Tropical Convergence Zone (ITCZ) which in turn has been shown to be sensitive to variations in Indian Ocean sea-surface temperatures. These sea surface temperatures vary from year to year, hence the onset, duration and intensity of seasonal rainfall varies considerably inter-annually.

Williams and Funk (2011) have recently suggested that the drying trend in parts of eastern Africa, Kenya and Sudan is related to a ~1°C warming of the sea surface temperature in the Indian Ocean.

Another, and well documented, ocean influence on rainfall in the region is ENSO . El Niño episodes usually cause greater than average rainfalls in the short rainy season (October, November and December, or OND), while cold phases(La Niña) are associated with drier than average seasons.

Some results that support the above conclusion of drying across much of the region are illustrated in Table 1 and Figure 1 from Tanzanian meteorological station data analysis. The slope column represents the millimeters change in annual rainfall. Arusha, for example, has experienced a decline of 3.2 mm per year since 1960. The significance column illustrates the statistical significance of the change. The stations in Figure 1 with red outline are those with statistically significant declines in rainfall. No station had statistically significant increase in  rainfall.

Rainfall Trends

Tanzania Station Annual Rainfall
Start Year End Year Slope Intercept Signif.
Arusha 1960 2010 -3.204 841.448 0.919 0.857
Buboka 1961 2005 -9.741 2238.295 0.993 1
DIA Dar es Salaam 1961 2005 -2.534 1206.313 0.955 0.577
Dodoma 1961 2005 1.317 539.283 0.954 0.68
Engare 1940 2010 -4.91 966.871 1.012 0.953
Iringa 1961 2005 -1.819 629.071 0.971 0.916
Karatu 1940 2010 -5.546 946.554 1.009 0.998
Kisangara 1940 2010 -0.261 644.791 0.969 0.402
Loliondo 1934 2008 -0.835 889.679 0.917 0.276
Lomwe 1960 2005 -10.822 1324.122 1.115 0.426
Lyamungu 1937 2008 -3.139 1634.869 0.957 0.904
Mbeya 1961 2005 -2.003 950.663 0.976 0.685
Monduli 1940 2010 -0.688 825.874 0.926 0.557
Mtwara 1961 2005 -5.89 1236.617 0.965 0.916
Musoma 1961 2005 -0.592 922.54 0.964 0.328
Same 1940 2010 -0.62 523.903 0.899 0.386
TPC Langasani 1940 2010 -0.186 422.647 0.774 0.642

Table 1. Trends in Annual Rainfall at Tanzanian Meteorological Stations, Statistical Analysis Results. Source: Olson et al. 2013.

Meteorological Stations in Tanzania

station-data

Figure 1 Source: Olson et al. 2013.

The graphs below in Figure 2 of annual rainfall in northern Tanzania from the 1930s to present illustrate some of these trends. The results point to the large variability in rainfall from year to year. In Same for example, the high inter-annual variability makes any 9-year average trends difficult to discern.

Click to enlarge.
Figure 2. Annual rainfall from meteorological stations in northern Tanzania.  Data source: Tanzanian Meteorological Agency.

Below in Figure 3 is a graph of daily rainfall data from Monduli in Northern Tanzania. It illustrates the seasonality of rainfall — from a short dry season in January and Febuary (day 1 to 50 or so), then the long rainy season, the long dry season, and finally the second, short rainy season in October to December.  Comparing the amount of rainfall between time periods (1935 to 1960 in black, 1961 to 1990 in red and 1991 to 2005 in blue), a trend becomes visible. Over time, the short rainy season has received less rainfall, the short dry season is dryer, and the rains during the long rainy season are ending earlier.  People depend especially on the long rainy season for the main crop harvest, yet it is becoming shorter. These changes in both rainy seasons have huge implications for agriculture.

monduli-avg-precip

Figure 3. Daily rainfall trends in Monduli, Tanzania from 1935 to 2005. Analysis source: Andresen et al. 2013. Data source: Tanzania Meteorological Agency.

Vegetation Trends

To better understand the impact of changes in temperature and rainfall, a statistical analysis of a vegetation was conducted across East Africa. The vegetation index used is the Normalized Difference Vegetation Index (NDVI) from GIMMS MODIS satellite imagery.  NDVI represents the amount of live, green vegetation as measured from satellite imagery. A new statistical procedure was developed to identify temporal trends, when they occurred and the shape of their trend (whether gradual, abrupt or like a hockey stick) (Kim et al. 2013).

sat-veg

Figure 4 shows imagery of vegetation production in East Africa. Source: Kim et al. 2013.

The results of the analysis indicate that:

  1. Significant declines in 25% of area
  2. Declines mostly south of equator, from 1970s
  3. Appear to be mostly climate-driven, warmer temperatures + declining or no change in precip.
  4. Increases in percent of vegetation north of equator, NE Kenya,  but vegetation amounts still low.

Temperature Trends

Temperature analyses shows a general increase of 1.5° to 2° C. across East Africa.

Summary of Trends of Current Climate in East Africa.

  1. Rainfall has remained steady or declined over most of the region during the past several decades. The average decreases during the past 50 years are on the order of 5-25%. The largest decreases in the region were found in southern Tanzania and central Kenya. In Tanzania, rainfall has decreased approx. 3 mm/month or 3% per decade.
  2. The decreases were in unimodal and in bimodal regions for both long and short seasons. Some of the decreases appear to be associated with a gradual compression or shortening of the rainy season (s).
  3. Long terms trends in western Uganda suggest increasing long season rainfall since 1960, although there have been decreases in that region during the most recent decade.
  4. In northeastern Kenya, rainfall and vegetation appear to have been increasing, especially in the past two decades, but there is uncertainty over these trends.
  5. Recent research suggests that the drying is related to a warming in the Indian Ocean. Given the persistence of sea surface temperatures, a continuation of recent trends is possible during the next decade or longer.
  6. Temperatures have risen approx. 0.2-0.4˚C/decade since 1960, which is inversely related to rainfall trends (warmer temperatures with less rainfall, fewer clouds).
  7. Rates of potential evapotranspiration and vegetation water needs have increased due to the warming temperatures. This has led to declining vegetation productivity across large areas.

Climate Trends observed by East African Farmers and Herders

Climate changes seen:

  1. Unreliable start of rainy seasons
  2. Breaks in mid-season
  3. Less rain overall
  4. Streams drying
  5. Too hot, too much sunshine.

Impacts experienced:

  1. Frequent harvest failure
  2. Lower yields
  3. More pests & diseases
  4. Less water for irrigation
  5. Big differences between men & women, rich & poor.

Expected impact of future climate change

  • Climate changes are seen as difficult to predict, but people expect climate variability to increase
  • Chronic water shortages if no appropriate interventions
  • Increased food insecurity
  • Fear and uncertainty in the land rights and use of future generations
  • Long sunshine hours (i.e. long dry seasons and droughts) in the future, and noted that without God’s divine intervention, the conditions will get worse.

References

Andresen, Jeff, Jennifer Olson, Philip K Thornton, Nathan Moore. 2013. Climate Change Impacts on Maize Production in East Africa. Gopal Alagarswamy, 04/2013; In proceeding of: The Association of American Geographers Annual Meeting.

Hoaglin, D.C., F. Mosteller and J.W. Tukey. 1983. Understanding robust and exploratory data analysis. Vol. 3. New York: Wiley.

Kim, Dong-Yun, Valerie Thomas, Jennifer Olson, Matthew Williams, Nicolle Clements. 2013. Statistical Trend and Change Point Analysis of Land Cover Change Patterns in East Africa. International Journal of Remote Sensing. 34(19):6636-6650. DOI:10.1080/01431161.2013.804224

Olson et al. 2013. Addressing the Impact of Climate Change on Agricultural Systems in East Africa. Presented at the University of Dar es Salaam and Michigan State University Climate Change Adaptation on Agriculture Sector Research Program, Research Policy Implications Workshop, Blue Pearl Hotel, Ubungo, Dar Es Salaam, Tanzania, March 1, 2013

Williams, A.P. and C. Funk. 2011. A westward extension of the warm pool leads to a westward extension of the Walker circulation, drying eastern Africa. Climate Dynamics. Vol. 37, pp 2417–2435