This paper describes how data from a variety of sources are merged to
present new countrywide maps of the solar energy distribution over Eth
iopia. The spatial coverage of stations with radiation data was found
to be unsatisfactory for the purpose of a countrywide solar energy ass
essment exercise. Therefore, radiation had to be predicted from sunshi
ne hours by employing empirical models. Using data from seven stations
in Ethiopia, linear and quadratic correlation relationships between m
onthly mean daily solar radiation and sunshine hours per day have been
developed. These regional models show a distinct improvement over pre
viously employed countrywide models. To produce a national solar-energ
y distribution profile, a spatial extension of the radiation/sunshine
relationships had to be carried out. To do this, the intercepts (a) an
d slopes (b) of each of the seven linear regression equations and anot
her six from previous studies, completed in neighbouring Sudan, Kenya
and Yemen, were used to interpolate the corresponding values to areas
between them. Subsequent to these procedures, 142 stations providing o
nly sunshine data were assigned their ''appropriate'' a and b values t
o estimate the amount of solar radiation received, which was then used
to produce annual and monthly solar radiation distribution maps for E
thiopia. The results show that in all regions solar energy is an abund
ant resource. Copyright (C) 1996 Elsevier Science Ltd.