The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B7. Beijing 2008 
appropriate interventions for different critical water uses, like 
irrigated agriculture, (Bastiaanssen et al., 2001). If reliable and 
consistent information regarding évapotranspiration can be 
accessed at low cost, it can be used to analyze the performance 
of for example irrigation systems and devise better management 
strategies. Remote sensing is one such an alternative, and as 
such, over the last decade, methods for calculating the actual 
évapotranspiration (ET a ) have been developed (Bastiaanssen, 
1995). The main advantage of this approach is that large areas 
are covered, and that data is easily obtainable without extensive 
monitoring networks in the field (Bandara, 2003; Chemin et al., 
2004). In summary the objectives of this study is to use satellite 
images to estimate spatio temporal variation of actual 
évapotranspiration across the basin from 2000 to 2006. 
2. THE STUDY AREA 
The Ewaso Ng’iro basin is the largest out of the five major 
drainage basins that make up the Kenyan drainage network, and 
covers 210,226km2 thus representing about 37% of the total 
area and contributes only about 7% of the total annual river 
flows. The basin constitutes the upper stream section of this 
drainage area, covering 15,251km2. It is situated between 
latitudes 0° 20’ south and 1° 01’ north and longitudes 36° 10’ 
east and 38° 00’ east as defined by the natural topographic 
divide. The Upper Ewaso Ng’iro basin drains from the Rift 
Valley escarpment to the west, the Nyandarua ranges to the 
north west, the Mt. Kenya to the south, the Nyambene hills in 
the east, and the Mathews ranges to the north while the 
downstream outlet is at Archers post. 
The eastern region of the basin has a clear bimodal distribution 
with highest rainfall in April and October (Berger, 1989) while 
in the western and north western region, beside the rainfall 
being bimodal pattern, continental rainfall is also experienced 
between June and September. The basin is a highland-lowland 
system and changes in elevation gives rise to a dramatic climate 
and ecological gradient, from humid moorlands and forests on 
the slopes to arid Acacia bushland in the lowland, with a diverse 
pattern of landuse (Decurtins, 1992). Most part of the basin is 
located in the leeward slopes of Mt. Kenya while the other part 
is in the windward slopes of the Nyandarua ranges, thus making 
it predominantly an ASAL region. Rainfall amounts are low, 
ranging from 2000 mm on the Nyandarua Ranges, to under 365 
mm per annum in the drier north eastern areas with mean annual 
rainfall of about 700 mm. Despite this relatively high amount of 
rainfall, the distribution is such that the seasonal amounts are 
insufficient for proper crop growth in most parts of the basin. 
3. MATERIALS AND METHODOLOGY 
MODIS LIB data which is an archive of 36 channels of visible 
and near-infrared reflectance and radiance, and thermal-infrared 
radiance was used. A total of 30 cloud free images covering the 
study area for 2000, 2003 and 2006, were downloaded from the 
satellite active archive’s website 
http://edcimswww.cr.usgs.gov/pub/imswelcome/. The images 
were processed to provide the necessary data required for the 
estimation of the areal évapotranspiration using the surface 
energy balance algorithm for land (SEBAL) approach. 
3.1 Surface Energy Balance Algorithm for Land (SEBAL) 
The Surface Energy Balance Algorithm for Land (SEBAL) is a 
relatively new parameterization of the energy balance and 
surface fluxes based on spectral satellite measurements. SEBAL 
requires spatially distributed Visible, Near-infrared and Thermal 
infrared input data, from satellite imageries. SEBAL 
parameterization is an iterative and feedback-based numerical 
procedure that deduces the radiation, heat and evaporative 
fluxes. The algorithm computes most essential hydro 
meteorological parameters and requires little field information 
(only incoming solar radiation, air temperature and wind speed 
data are required), (Bastiaanssen, 1998a &b, 2000). The energy 
balance during the satellite overpass and the integrated 24HRS 
fluxes are computed on pixel by pixel basis. 
The model comprises of a number of computational steps for 
image processing and finally calculates the actual 
évapotranspiration (ET a ) as well as other energy exchanges 
between land and atmosphere. By ignoring energy required for 
photosynthesis and the heat storage in vegetation, and in its 
most simplified form SEBAL reads as: R n = G 0 + H + LE where 
R n is the net radiation absorbed at the land surface (W/m2), G 0 
the soil heat flux to warm or cool the soil (W/m2), H the 
sensible heat flux to warm or cool the atmosphere (W/m2) and 
LE is the latent heat flux associated with evaporation of water 
from soil, water and vegetation (W/m2). The applied SEBAL 
method consists of a physically based one-layer sensible heat 
transfer scheme and an empirical estimation scheme for soil 
heat flux. The soil heat flux is computed as an empirical fraction 
of the net radiation using surface temperature, surface albedo 
and the normalized vegetation index (NDVI), as the depending 
variables, see figure below. 
jUrM «atSKv Sort aux Moêsfcttv 
SLUAL 
Figure 3: Conceptual Scheme for SEBAL showing its principal 
Components, Bastiaanssen et al, 1998. 
3.2 Estimating Energy Balance Components 
The surface energy balance is expressed as follows: 
R n = H + LE + G 0 (i) 
Figure 2: The Upper Ewaso Ng’iro Baisn